{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9988751406074241,
  "eval_steps": 500,
  "global_step": 666,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "clip_ratio": 0.0,
      "completion_length": 150.88541984558105,
      "epoch": 0.0014998125234345708,
      "grad_norm": 2.0143966719522566,
      "kl": 0.0,
      "learning_rate": 2.9850746268656716e-07,
      "loss": -0.123,
      "reward": 0.39659371972084045,
      "reward_std": 0.6371014267206192,
      "rewards/accuracy_reward": 0.1250000037252903,
      "rewards/reasoning_steps_reward": 0.04861111380159855,
      "rewards/repetition_penalty_reward": -0.05305909365415573,
      "rewards/tag_count_reward": 0.2760416716337204,
      "step": 1
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 139.43750381469727,
      "epoch": 0.0029996250468691415,
      "grad_norm": 2.260545366465926,
      "kl": 0.0,
      "learning_rate": 5.970149253731343e-07,
      "loss": -0.0563,
      "reward": 0.29565349593758583,
      "reward_std": 0.5291136056184769,
      "rewards/accuracy_reward": 0.07812500046566129,
      "rewards/reasoning_steps_reward": 0.039930558297783136,
      "rewards/repetition_penalty_reward": -0.04505832493305206,
      "rewards/tag_count_reward": 0.2226562537252903,
      "step": 2
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 143.8854217529297,
      "epoch": 0.0044994375703037125,
      "grad_norm": 2.04474730883385,
      "kl": 0.0003943443298339844,
      "learning_rate": 8.955223880597015e-07,
      "loss": -0.128,
      "reward": 0.41914862394332886,
      "reward_std": 0.6219020187854767,
      "rewards/accuracy_reward": 0.13020833861082792,
      "rewards/reasoning_steps_reward": 0.06423611287027597,
      "rewards/repetition_penalty_reward": -0.05524375103414059,
      "rewards/tag_count_reward": 0.2799479216337204,
      "step": 3
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 143.52605056762695,
      "epoch": 0.005999250093738283,
      "grad_norm": 2.088074149444087,
      "kl": 0.0003809928894042969,
      "learning_rate": 1.1940298507462686e-06,
      "loss": -0.055,
      "reward": 0.32894811406731606,
      "reward_std": 0.5705550760030746,
      "rewards/accuracy_reward": 0.08333333628252149,
      "rewards/reasoning_steps_reward": 0.05381944729015231,
      "rewards/repetition_penalty_reward": -0.03997551556676626,
      "rewards/tag_count_reward": 0.2317708395421505,
      "step": 4
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 131.28125381469727,
      "epoch": 0.0074990626171728535,
      "grad_norm": 2.0380845412513424,
      "kl": 0.0005826950073242188,
      "learning_rate": 1.4925373134328358e-06,
      "loss": -0.066,
      "reward": 0.2540438659489155,
      "reward_std": 0.45853011310100555,
      "rewards/accuracy_reward": 0.05208333395421505,
      "rewards/reasoning_steps_reward": 0.04687500232830644,
      "rewards/repetition_penalty_reward": -0.04413322079926729,
      "rewards/tag_count_reward": 0.19921875651925802,
      "step": 5
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 154.97917556762695,
      "epoch": 0.008998875140607425,
      "grad_norm": 1.891201241821577,
      "kl": 0.0010385513305664062,
      "learning_rate": 1.791044776119403e-06,
      "loss": -0.0654,
      "reward": 0.45939914882183075,
      "reward_std": 0.5917806774377823,
      "rewards/accuracy_reward": 0.1093750037252903,
      "rewards/reasoning_steps_reward": 0.059027780778706074,
      "rewards/repetition_penalty_reward": -0.05145154893398285,
      "rewards/tag_count_reward": 0.3424479216337204,
      "step": 6
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 135.93229484558105,
      "epoch": 0.010498687664041995,
      "grad_norm": 2.1855192003244244,
      "kl": 0.00994873046875,
      "learning_rate": 2.08955223880597e-06,
      "loss": -0.0289,
      "reward": 0.47621314972639084,
      "reward_std": 0.6438451856374741,
      "rewards/accuracy_reward": 0.13020833861082792,
      "rewards/reasoning_steps_reward": 0.03819444729015231,
      "rewards/repetition_penalty_reward": -0.05026257690042257,
      "rewards/tag_count_reward": 0.358072929084301,
      "step": 7
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 149.20313262939453,
      "epoch": 0.011998500187476566,
      "grad_norm": 1.5337300270924847,
      "kl": 0.027801513671875,
      "learning_rate": 2.3880597014925373e-06,
      "loss": -0.0828,
      "reward": 0.5498909652233124,
      "reward_std": 0.6071374714374542,
      "rewards/accuracy_reward": 0.1354166716337204,
      "rewards/reasoning_steps_reward": 0.0763888917863369,
      "rewards/repetition_penalty_reward": -0.05123751983046532,
      "rewards/tag_count_reward": 0.3893229365348816,
      "step": 8
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 163.7291717529297,
      "epoch": 0.013498312710911136,
      "grad_norm": 27.976622447462567,
      "kl": 0.797607421875,
      "learning_rate": 2.686567164179105e-06,
      "loss": -0.0138,
      "reward": 0.8464770168066025,
      "reward_std": 0.7221487462520599,
      "rewards/accuracy_reward": 0.2708333469927311,
      "rewards/reasoning_steps_reward": 0.09895833767950535,
      "rewards/repetition_penalty_reward": -0.06498134508728981,
      "rewards/tag_count_reward": 0.541666679084301,
      "step": 9
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 166.1145896911621,
      "epoch": 0.014998125234345707,
      "grad_norm": 173.04994011488648,
      "kl": 4.0771484375,
      "learning_rate": 2.9850746268656716e-06,
      "loss": 0.0723,
      "reward": 0.9919786900281906,
      "reward_std": 0.7005721777677536,
      "rewards/accuracy_reward": 0.3593750074505806,
      "rewards/reasoning_steps_reward": 0.07291667279787362,
      "rewards/repetition_penalty_reward": -0.08093799650669098,
      "rewards/tag_count_reward": 0.6406250149011612,
      "step": 10
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 163.34375381469727,
      "epoch": 0.016497937757780277,
      "grad_norm": 745.8172627492539,
      "kl": 13.115234375,
      "learning_rate": 3.283582089552239e-06,
      "loss": 0.2316,
      "reward": 1.0233316719532013,
      "reward_std": 0.7426625639200211,
      "rewards/accuracy_reward": 0.385416679084301,
      "rewards/reasoning_steps_reward": 0.045138892251998186,
      "rewards/repetition_penalty_reward": -0.07649472542107105,
      "rewards/tag_count_reward": 0.669270858168602,
      "step": 11
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 168.0729217529297,
      "epoch": 0.01799775028121485,
      "grad_norm": 9.77279564737439,
      "kl": 0.47802734375,
      "learning_rate": 3.582089552238806e-06,
      "loss": -0.0793,
      "reward": 1.0776500403881073,
      "reward_std": 0.6928116679191589,
      "rewards/accuracy_reward": 0.354166679084301,
      "rewards/reasoning_steps_reward": 0.10243056155741215,
      "rewards/repetition_penalty_reward": -0.06774928979575634,
      "rewards/tag_count_reward": 0.688802108168602,
      "step": 12
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 150.58855056762695,
      "epoch": 0.01949756280464942,
      "grad_norm": 1.9563170638007599,
      "kl": 0.1324462890625,
      "learning_rate": 3.8805970149253735e-06,
      "loss": -0.005,
      "reward": 1.2495136260986328,
      "reward_std": 0.6537359356880188,
      "rewards/accuracy_reward": 0.4270833432674408,
      "rewards/reasoning_steps_reward": 0.0538194477558136,
      "rewards/repetition_penalty_reward": -0.051701731979846954,
      "rewards/tag_count_reward": 0.8203125149011612,
      "step": 13
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 144.01042366027832,
      "epoch": 0.02099737532808399,
      "grad_norm": 7.69562170272425,
      "kl": 0.294189453125,
      "learning_rate": 4.17910447761194e-06,
      "loss": 0.0198,
      "reward": 1.16742305457592,
      "reward_std": 0.6400385946035385,
      "rewards/accuracy_reward": 0.354166679084301,
      "rewards/reasoning_steps_reward": 0.05902778171002865,
      "rewards/repetition_penalty_reward": -0.062177615240216255,
      "rewards/tag_count_reward": 0.8164062798023224,
      "step": 14
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 156.21875762939453,
      "epoch": 0.02249718785151856,
      "grad_norm": 2.534410431741499,
      "kl": 0.146728515625,
      "learning_rate": 4.477611940298508e-06,
      "loss": 0.0501,
      "reward": 1.4073261320590973,
      "reward_std": 0.645281046628952,
      "rewards/accuracy_reward": 0.494791679084301,
      "rewards/reasoning_steps_reward": 0.1354166716337204,
      "rewards/repetition_penalty_reward": -0.044496800750494,
      "rewards/tag_count_reward": 0.8216145932674408,
      "step": 15
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 142.89062881469727,
      "epoch": 0.023997000374953132,
      "grad_norm": 1.6522024022194932,
      "kl": 0.08740234375,
      "learning_rate": 4.7761194029850745e-06,
      "loss": 0.004,
      "reward": 1.378474086523056,
      "reward_std": 0.6564009487628937,
      "rewards/accuracy_reward": 0.4687500298023224,
      "rewards/reasoning_steps_reward": 0.1215277798473835,
      "rewards/repetition_penalty_reward": -0.04513707011938095,
      "rewards/tag_count_reward": 0.8333333432674408,
      "step": 16
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 160.17708587646484,
      "epoch": 0.0254968128983877,
      "grad_norm": 1.2954290652113556,
      "kl": 0.07891845703125,
      "learning_rate": 5.074626865671642e-06,
      "loss": 0.0088,
      "reward": 1.3321838080883026,
      "reward_std": 0.6731031388044357,
      "rewards/accuracy_reward": 0.421875,
      "rewards/reasoning_steps_reward": 0.1371527872979641,
      "rewards/repetition_penalty_reward": -0.04976071882992983,
      "rewards/tag_count_reward": 0.8229166865348816,
      "step": 17
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 156.93750381469727,
      "epoch": 0.02699662542182227,
      "grad_norm": 1.3681933966105373,
      "kl": 0.05126953125,
      "learning_rate": 5.37313432835821e-06,
      "loss": -0.0754,
      "reward": 1.3372644186019897,
      "reward_std": 0.6545021533966064,
      "rewards/accuracy_reward": 0.401041679084301,
      "rewards/reasoning_steps_reward": 0.2013889029622078,
      "rewards/repetition_penalty_reward": -0.062041101045906544,
      "rewards/tag_count_reward": 0.7968750149011612,
      "step": 18
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 175.35937881469727,
      "epoch": 0.028496437945256844,
      "grad_norm": 1.1392001469038255,
      "kl": 0.05596923828125,
      "learning_rate": 5.671641791044776e-06,
      "loss": -0.0216,
      "reward": 1.352351814508438,
      "reward_std": 0.6970622688531876,
      "rewards/accuracy_reward": 0.4375000074505806,
      "rewards/reasoning_steps_reward": 0.2031250074505806,
      "rewards/repetition_penalty_reward": -0.06561699230223894,
      "rewards/tag_count_reward": 0.7773437649011612,
      "step": 19
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 194.78646087646484,
      "epoch": 0.029996250468691414,
      "grad_norm": 1.4083775545206836,
      "kl": 0.0611572265625,
      "learning_rate": 5.970149253731343e-06,
      "loss": -0.1554,
      "reward": 1.4778603315353394,
      "reward_std": 0.6467646211385727,
      "rewards/accuracy_reward": 0.4375,
      "rewards/reasoning_steps_reward": 0.2829861342906952,
      "rewards/repetition_penalty_reward": -0.06684454903006554,
      "rewards/tag_count_reward": 0.8242187798023224,
      "step": 20
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 221.3541717529297,
      "epoch": 0.031496062992125984,
      "grad_norm": 0.9468941237796497,
      "kl": 0.050048828125,
      "learning_rate": 6.2686567164179116e-06,
      "loss": -0.1031,
      "reward": 1.576261043548584,
      "reward_std": 0.7107058763504028,
      "rewards/accuracy_reward": 0.4427083432674408,
      "rewards/reasoning_steps_reward": 0.4444444701075554,
      "rewards/repetition_penalty_reward": -0.0830271951854229,
      "rewards/tag_count_reward": 0.7721354365348816,
      "step": 21
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 253.48438262939453,
      "epoch": 0.032995875515560553,
      "grad_norm": 0.8523009396371123,
      "kl": 0.05426025390625,
      "learning_rate": 6.567164179104478e-06,
      "loss": -0.0433,
      "reward": 1.9110043048858643,
      "reward_std": 0.7316416054964066,
      "rewards/accuracy_reward": 0.5416666865348816,
      "rewards/reasoning_steps_reward": 0.6388888955116272,
      "rewards/repetition_penalty_reward": -0.08595757372677326,
      "rewards/tag_count_reward": 0.8164062798023224,
      "step": 22
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 282.4323043823242,
      "epoch": 0.03449568803899512,
      "grad_norm": 0.8161516191732155,
      "kl": 0.07037353515625,
      "learning_rate": 6.865671641791045e-06,
      "loss": -0.0343,
      "reward": 2.085984379053116,
      "reward_std": 0.670979842543602,
      "rewards/accuracy_reward": 0.5625000298023224,
      "rewards/reasoning_steps_reward": 0.7743056118488312,
      "rewards/repetition_penalty_reward": -0.10238376632332802,
      "rewards/tag_count_reward": 0.8515625149011612,
      "step": 23
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 277.39064025878906,
      "epoch": 0.0359955005624297,
      "grad_norm": 0.88714678428363,
      "kl": 0.08349609375,
      "learning_rate": 7.164179104477612e-06,
      "loss": -0.0093,
      "reward": 2.1416602730751038,
      "reward_std": 0.6270937323570251,
      "rewards/accuracy_reward": 0.5416666865348816,
      "rewards/reasoning_steps_reward": 0.8420138955116272,
      "rewards/repetition_penalty_reward": -0.10530160553753376,
      "rewards/tag_count_reward": 0.8632812798023224,
      "step": 24
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 303.7708435058594,
      "epoch": 0.03749531308586427,
      "grad_norm": 0.7677870229716102,
      "kl": 0.0814208984375,
      "learning_rate": 7.46268656716418e-06,
      "loss": -0.0311,
      "reward": 2.169845759868622,
      "reward_std": 0.5908740907907486,
      "rewards/accuracy_reward": 0.4843750149011612,
      "rewards/reasoning_steps_reward": 0.8958334028720856,
      "rewards/repetition_penalty_reward": -0.119216812774539,
      "rewards/tag_count_reward": 0.9088542014360428,
      "step": 25
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 300.0052185058594,
      "epoch": 0.03899512560929884,
      "grad_norm": 0.751193936667339,
      "kl": 0.084228515625,
      "learning_rate": 7.761194029850747e-06,
      "loss": 0.0118,
      "reward": 2.2213982343673706,
      "reward_std": 0.5355894565582275,
      "rewards/accuracy_reward": 0.5000000298023224,
      "rewards/reasoning_steps_reward": 0.9461806118488312,
      "rewards/repetition_penalty_reward": -0.1258239597082138,
      "rewards/tag_count_reward": 0.9010416716337204,
      "step": 26
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 294.46875762939453,
      "epoch": 0.04049493813273341,
      "grad_norm": 0.817822463608519,
      "kl": 0.0811767578125,
      "learning_rate": 8.059701492537314e-06,
      "loss": 0.0218,
      "reward": 2.4470449686050415,
      "reward_std": 0.5053400099277496,
      "rewards/accuracy_reward": 0.6145833432674408,
      "rewards/reasoning_steps_reward": 0.9843750298023224,
      "rewards/repetition_penalty_reward": -0.11154878698289394,
      "rewards/tag_count_reward": 0.9596354365348816,
      "step": 27
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 285.4166717529297,
      "epoch": 0.04199475065616798,
      "grad_norm": 0.8329974091920429,
      "kl": 0.101806640625,
      "learning_rate": 8.35820895522388e-06,
      "loss": 0.0405,
      "reward": 2.4431938529014587,
      "reward_std": 0.46228019893169403,
      "rewards/accuracy_reward": 0.614583358168602,
      "rewards/reasoning_steps_reward": 0.9930555671453476,
      "rewards/repetition_penalty_reward": -0.10975776612758636,
      "rewards/tag_count_reward": 0.9453125298023224,
      "step": 28
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 251.47396850585938,
      "epoch": 0.04349456317960255,
      "grad_norm": 2.1092410850036143,
      "kl": 0.1441650390625,
      "learning_rate": 8.656716417910447e-06,
      "loss": 0.0865,
      "reward": 2.475065588951111,
      "reward_std": 0.47668417543172836,
      "rewards/accuracy_reward": 0.661458358168602,
      "rewards/reasoning_steps_reward": 0.9947916716337204,
      "rewards/repetition_penalty_reward": -0.1121740210801363,
      "rewards/tag_count_reward": 0.9309895932674408,
      "step": 29
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 263.27083587646484,
      "epoch": 0.04499437570303712,
      "grad_norm": 0.8134902475342992,
      "kl": 0.140380859375,
      "learning_rate": 8.955223880597016e-06,
      "loss": 0.0415,
      "reward": 2.441520571708679,
      "reward_std": 0.3973531872034073,
      "rewards/accuracy_reward": 0.6093750223517418,
      "rewards/reasoning_steps_reward": 0.9982638955116272,
      "rewards/repetition_penalty_reward": -0.1179413478821516,
      "rewards/tag_count_reward": 0.9518229365348816,
      "step": 30
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 230.7916717529297,
      "epoch": 0.046494188226471694,
      "grad_norm": 0.8575684864473243,
      "kl": 0.1580810546875,
      "learning_rate": 9.253731343283582e-06,
      "loss": 0.0852,
      "reward": 2.5657079219818115,
      "reward_std": 0.4555797800421715,
      "rewards/accuracy_reward": 0.708333358168602,
      "rewards/reasoning_steps_reward": 1.0,
      "rewards/repetition_penalty_reward": -0.09314640611410141,
      "rewards/tag_count_reward": 0.9505208432674408,
      "step": 31
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 208.65625762939453,
      "epoch": 0.047994000749906264,
      "grad_norm": 1.2887250609562435,
      "kl": 0.23876953125,
      "learning_rate": 9.552238805970149e-06,
      "loss": 0.0936,
      "reward": 2.6088263988494873,
      "reward_std": 0.3428891524672508,
      "rewards/accuracy_reward": 0.755208358168602,
      "rewards/reasoning_steps_reward": 0.9982638955116272,
      "rewards/repetition_penalty_reward": -0.06912498734891415,
      "rewards/tag_count_reward": 0.9244791865348816,
      "step": 32
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 195.6197967529297,
      "epoch": 0.049493813273340834,
      "grad_norm": 0.998549892430521,
      "kl": 0.164306640625,
      "learning_rate": 9.850746268656717e-06,
      "loss": 0.0716,
      "reward": 2.281342178583145,
      "reward_std": 0.3404741808772087,
      "rewards/accuracy_reward": 0.4062500186264515,
      "rewards/reasoning_steps_reward": 0.9965277910232544,
      "rewards/repetition_penalty_reward": -0.06414407771080732,
      "rewards/tag_count_reward": 0.942708358168602,
      "step": 33
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 173.95312881469727,
      "epoch": 0.0509936257967754,
      "grad_norm": 1.317082973370953,
      "kl": 0.213134765625,
      "learning_rate": 1.0149253731343284e-05,
      "loss": 0.1399,
      "reward": 2.370617628097534,
      "reward_std": 0.4167383909225464,
      "rewards/accuracy_reward": 0.5260416716337204,
      "rewards/reasoning_steps_reward": 0.9878472536802292,
      "rewards/repetition_penalty_reward": -0.0443130349740386,
      "rewards/tag_count_reward": 0.9010416865348816,
      "step": 34
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 172.42187881469727,
      "epoch": 0.05249343832020997,
      "grad_norm": 1.1847166733741281,
      "kl": 0.202392578125,
      "learning_rate": 1.0447761194029851e-05,
      "loss": 0.1261,
      "reward": 2.3955256938934326,
      "reward_std": 0.328448873013258,
      "rewards/accuracy_reward": 0.5312500149011612,
      "rewards/reasoning_steps_reward": 0.97743059694767,
      "rewards/repetition_penalty_reward": -0.04284239187836647,
      "rewards/tag_count_reward": 0.9296875149011612,
      "step": 35
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 193.6458396911621,
      "epoch": 0.05399325084364454,
      "grad_norm": 1.4752348010664043,
      "kl": 0.24560546875,
      "learning_rate": 1.074626865671642e-05,
      "loss": 0.4191,
      "reward": 2.3304308652877808,
      "reward_std": 0.5351479351520538,
      "rewards/accuracy_reward": 0.5520833507180214,
      "rewards/reasoning_steps_reward": 0.9340278208255768,
      "rewards/repetition_penalty_reward": -0.0515137268230319,
      "rewards/tag_count_reward": 0.8958333432674408,
      "step": 36
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 182.2135467529297,
      "epoch": 0.05549306336707911,
      "grad_norm": 1.7682892736323388,
      "kl": 0.32080078125,
      "learning_rate": 1.1044776119402986e-05,
      "loss": 0.402,
      "reward": 2.4258365631103516,
      "reward_std": 0.5576556175947189,
      "rewards/accuracy_reward": 0.6562500149011612,
      "rewards/reasoning_steps_reward": 0.89930559694767,
      "rewards/repetition_penalty_reward": -0.038573198951780796,
      "rewards/tag_count_reward": 0.9088541716337204,
      "step": 37
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 240.62500381469727,
      "epoch": 0.05699287589051369,
      "grad_norm": 1.8853983944936439,
      "kl": 0.3203125,
      "learning_rate": 1.1343283582089553e-05,
      "loss": 0.655,
      "reward": 2.015189290046692,
      "reward_std": 0.6033422723412514,
      "rewards/accuracy_reward": 0.3697916716337204,
      "rewards/reasoning_steps_reward": 0.8541667312383652,
      "rewards/repetition_penalty_reward": -0.055123341269791126,
      "rewards/tag_count_reward": 0.8463541865348816,
      "step": 38
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 295.5833396911621,
      "epoch": 0.05849268841394826,
      "grad_norm": 3.2310358216677733,
      "kl": 0.3486328125,
      "learning_rate": 1.1641791044776121e-05,
      "loss": 0.7056,
      "reward": 2.03124737739563,
      "reward_std": 0.8326994776725769,
      "rewards/accuracy_reward": 0.5052083432674408,
      "rewards/reasoning_steps_reward": 0.7847221940755844,
      "rewards/repetition_penalty_reward": -0.052954130340367556,
      "rewards/tag_count_reward": 0.794270858168602,
      "step": 39
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 344.5208511352539,
      "epoch": 0.05999250093738283,
      "grad_norm": 38.31384293221201,
      "kl": 0.46484375,
      "learning_rate": 1.1940298507462686e-05,
      "loss": 0.6778,
      "reward": 1.9736978709697723,
      "reward_std": 0.8218565732240677,
      "rewards/accuracy_reward": 0.4843750149011612,
      "rewards/reasoning_steps_reward": 0.7552083283662796,
      "rewards/repetition_penalty_reward": -0.013281408930197358,
      "rewards/tag_count_reward": 0.747395858168602,
      "step": 40
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 443.59375762939453,
      "epoch": 0.0614923134608174,
      "grad_norm": 10963112.509325096,
      "kl": 38711.625,
      "learning_rate": 1.2238805970149255e-05,
      "loss": 3011.3784,
      "reward": 1.5866824984550476,
      "reward_std": 0.7351708710193634,
      "rewards/accuracy_reward": 0.3020833395421505,
      "rewards/reasoning_steps_reward": 0.6545138955116272,
      "rewards/repetition_penalty_reward": -0.01444607856683433,
      "rewards/tag_count_reward": 0.6445312798023224,
      "step": 41
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 571.9479370117188,
      "epoch": 0.06299212598425197,
      "grad_norm": 1562.2702573171862,
      "kl": 13.8759765625,
      "learning_rate": 1.2537313432835823e-05,
      "loss": 1.2646,
      "reward": 1.396679848432541,
      "reward_std": 0.764959841966629,
      "rewards/accuracy_reward": 0.2604166753590107,
      "rewards/reasoning_steps_reward": 0.5763889253139496,
      "rewards/repetition_penalty_reward": -0.006532001192681491,
      "rewards/tag_count_reward": 0.5664062649011612,
      "step": 42
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 818.1614837646484,
      "epoch": 0.06449193850768654,
      "grad_norm": 36.12141290169206,
      "kl": 1.1787109375,
      "learning_rate": 1.2835820895522388e-05,
      "loss": 0.1815,
      "reward": 1.0434824973344803,
      "reward_std": 0.6087194085121155,
      "rewards/accuracy_reward": 0.07812500279396772,
      "rewards/reasoning_steps_reward": 0.5711805820465088,
      "rewards/repetition_penalty_reward": -0.002958516124635935,
      "rewards/tag_count_reward": 0.397135429084301,
      "step": 43
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 883.0052185058594,
      "epoch": 0.06599175103112111,
      "grad_norm": 12.05770882883392,
      "kl": 1.087890625,
      "learning_rate": 1.3134328358208957e-05,
      "loss": 0.1592,
      "reward": 0.8965069055557251,
      "reward_std": 0.5693257004022598,
      "rewards/accuracy_reward": 0.057291668839752674,
      "rewards/reasoning_steps_reward": 0.508680559694767,
      "rewards/repetition_penalty_reward": -0.0014966219605412334,
      "rewards/tag_count_reward": 0.3320312574505806,
      "step": 44
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 978.4687805175781,
      "epoch": 0.06749156355455568,
      "grad_norm": 7.679672597113327,
      "kl": 1.017578125,
      "learning_rate": 1.3432835820895525e-05,
      "loss": 0.0491,
      "reward": 0.7406667172908783,
      "reward_std": 0.41688016057014465,
      "rewards/accuracy_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.4843750223517418,
      "rewards/repetition_penalty_reward": -0.0002187521276937332,
      "rewards/tag_count_reward": 0.2565104253590107,
      "step": 45
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 992.9739837646484,
      "epoch": 0.06899137607799025,
      "grad_norm": 3776.078403846416,
      "kl": 122.390625,
      "learning_rate": 1.373134328358209e-05,
      "loss": 5.2459,
      "reward": 0.6156022548675537,
      "reward_std": 0.45898766070604324,
      "rewards/accuracy_reward": 0.0052083334885537624,
      "rewards/reasoning_steps_reward": 0.3975694626569748,
      "rewards/repetition_penalty_reward": -0.0007172302284743637,
      "rewards/tag_count_reward": 0.2135416716337204,
      "step": 46
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 1021.09375,
      "epoch": 0.07049118860142482,
      "grad_norm": 53.535086229947574,
      "kl": 2.7412109375,
      "learning_rate": 1.4029850746268658e-05,
      "loss": 0.1119,
      "reward": 0.6543789356946945,
      "reward_std": 0.39667317271232605,
      "rewards/accuracy_reward": 0.0052083334885537624,
      "rewards/reasoning_steps_reward": 0.4670139029622078,
      "rewards/repetition_penalty_reward": -0.00013500768363883253,
      "rewards/tag_count_reward": 0.1822916716337204,
      "step": 47
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 1006.2812652587891,
      "epoch": 0.0719910011248594,
      "grad_norm": 9.182715345455021,
      "kl": 0.677734375,
      "learning_rate": 1.4328358208955224e-05,
      "loss": 0.0131,
      "reward": 0.5155243277549744,
      "reward_std": 0.3933122977614403,
      "rewards/accuracy_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.3645833507180214,
      "rewards/repetition_penalty_reward": -0.00010070096323033795,
      "rewards/tag_count_reward": 0.1510416716337204,
      "step": 48
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 1017.1927185058594,
      "epoch": 0.07349081364829396,
      "grad_norm": 2.136051505149193,
      "kl": 0.76416015625,
      "learning_rate": 1.4626865671641792e-05,
      "loss": 0.0258,
      "reward": 0.5431747883558273,
      "reward_std": 0.3834214210510254,
      "rewards/accuracy_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.3871527835726738,
      "rewards/repetition_penalty_reward": -0.0002280306780448882,
      "rewards/tag_count_reward": 0.1562500037252903,
      "step": 49
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 1005.1770935058594,
      "epoch": 0.07499062617172854,
      "grad_norm": 233.5775916802967,
      "kl": 15.7265625,
      "learning_rate": 1.492537313432836e-05,
      "loss": 0.6288,
      "reward": 0.5613471269607544,
      "reward_std": 0.40113527327775955,
      "rewards/accuracy_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.418402798473835,
      "rewards/repetition_penalty_reward": -0.0002848635776899755,
      "rewards/tag_count_reward": 0.1432291716337204,
      "step": 50
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 994.8333435058594,
      "epoch": 0.0764904386951631,
      "grad_norm": 10.304724683124707,
      "kl": 1.4677734375,
      "learning_rate": 1.5223880597014925e-05,
      "loss": 0.0407,
      "reward": 0.5884975641965866,
      "reward_std": 0.4051181599497795,
      "rewards/accuracy_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.4427083507180214,
      "rewards/repetition_penalty_reward": -4.4144707317173015e-05,
      "rewards/tag_count_reward": 0.1458333395421505,
      "step": 51
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 983.5208435058594,
      "epoch": 0.07799025121859768,
      "grad_norm": 56.749232087158944,
      "kl": 6.04296875,
      "learning_rate": 1.5522388059701494e-05,
      "loss": 0.2016,
      "reward": 0.6742116063833237,
      "reward_std": 0.4244392439723015,
      "rewards/accuracy_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.5104166716337204,
      "rewards/repetition_penalty_reward": -0.0002676011572475545,
      "rewards/tag_count_reward": 0.1640625037252903,
      "step": 52
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 972.0052185058594,
      "epoch": 0.07949006374203224,
      "grad_norm": 5.494493067960978,
      "kl": 2.2734375,
      "learning_rate": 1.582089552238806e-05,
      "loss": 0.0395,
      "reward": 0.6165060997009277,
      "reward_std": 0.412445493042469,
      "rewards/accuracy_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.4565972536802292,
      "rewards/repetition_penalty_reward": -0.00024740799563005567,
      "rewards/tag_count_reward": 0.1601562537252903,
      "step": 53
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 997.3750152587891,
      "epoch": 0.08098987626546682,
      "grad_norm": 8.732055316405138,
      "kl": 0.92333984375,
      "learning_rate": 1.6119402985074627e-05,
      "loss": 0.0115,
      "reward": 0.8314266949892044,
      "reward_std": 0.3945605829358101,
      "rewards/accuracy_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.6649305820465088,
      "rewards/repetition_penalty_reward": -0.00017056526121450588,
      "rewards/tag_count_reward": 0.1666666716337204,
      "step": 54
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 961.9583587646484,
      "epoch": 0.0824896887889014,
      "grad_norm": 14.31298550812026,
      "kl": 1.4169921875,
      "learning_rate": 1.6417910447761197e-05,
      "loss": -0.0265,
      "reward": 0.8787982016801834,
      "reward_std": 0.3934568166732788,
      "rewards/accuracy_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.7152778208255768,
      "rewards/repetition_penalty_reward": -0.0005421093846962322,
      "rewards/tag_count_reward": 0.1640625037252903,
      "step": 55
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 974.8125152587891,
      "epoch": 0.08398950131233596,
      "grad_norm": 262.24387888017645,
      "kl": 24.07275390625,
      "learning_rate": 1.671641791044776e-05,
      "loss": 0.9471,
      "reward": 0.975603461265564,
      "reward_std": 0.3715866580605507,
      "rewards/accuracy_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.7951389253139496,
      "rewards/repetition_penalty_reward": -0.0005250562171568163,
      "rewards/tag_count_reward": 0.1809895895421505,
      "step": 56
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 975.3854370117188,
      "epoch": 0.08548931383577053,
      "grad_norm": 38.16302080966472,
      "kl": 6.939453125,
      "learning_rate": 1.701492537313433e-05,
      "loss": 0.1715,
      "reward": 1.1071399450302124,
      "reward_std": 0.2679474614560604,
      "rewards/accuracy_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.9114583730697632,
      "rewards/repetition_penalty_reward": -0.0009330366592621431,
      "rewards/tag_count_reward": 0.1966145858168602,
      "step": 57
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 981.3020935058594,
      "epoch": 0.0869891263592051,
      "grad_norm": 2.325182725021432,
      "kl": 0.68359375,
      "learning_rate": 1.7313432835820894e-05,
      "loss": -0.0424,
      "reward": 1.130223572254181,
      "reward_std": 0.20110024139285088,
      "rewards/accuracy_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.9496527910232544,
      "rewards/repetition_penalty_reward": -0.003022975695785135,
      "rewards/tag_count_reward": 0.1835937574505806,
      "step": 58
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 1004.2187652587891,
      "epoch": 0.08848893888263967,
      "grad_norm": 2.918231849637039,
      "kl": 1.49267578125,
      "learning_rate": 1.7611940298507464e-05,
      "loss": 0.0337,
      "reward": 1.1544174253940582,
      "reward_std": 0.20146211609244347,
      "rewards/accuracy_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.9565972536802292,
      "rewards/repetition_penalty_reward": -0.014419377315789461,
      "rewards/tag_count_reward": 0.2122395895421505,
      "step": 59
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 1005.3697967529297,
      "epoch": 0.08998875140607424,
      "grad_norm": 1.073898600293748,
      "kl": 0.59765625,
      "learning_rate": 1.791044776119403e-05,
      "loss": -0.0304,
      "reward": 1.1241816580295563,
      "reward_std": 0.17474635317921638,
      "rewards/accuracy_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.9739583432674408,
      "rewards/repetition_penalty_reward": -0.06592248193919659,
      "rewards/tag_count_reward": 0.2161458358168602,
      "step": 60
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 1009.3020935058594,
      "epoch": 0.09148856392950881,
      "grad_norm": 139.56708057657698,
      "kl": 16.713623046875,
      "learning_rate": 1.8208955223880598e-05,
      "loss": 0.64,
      "reward": 1.0292470455169678,
      "reward_std": 0.155818872153759,
      "rewards/accuracy_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.9791666865348816,
      "rewards/repetition_penalty_reward": -0.17908628657460213,
      "rewards/tag_count_reward": 0.2291666753590107,
      "step": 61
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 1006.1718902587891,
      "epoch": 0.09298837645294339,
      "grad_norm": 0.6239077670049449,
      "kl": 0.2763671875,
      "learning_rate": 1.8507462686567165e-05,
      "loss": -0.0267,
      "reward": 0.9351680278778076,
      "reward_std": 0.12725192122161388,
      "rewards/accuracy_reward": 0.0,
      "rewards/reasoning_steps_reward": 0.9861111342906952,
      "rewards/repetition_penalty_reward": -0.2918284982442856,
      "rewards/tag_count_reward": 0.2408854179084301,
      "step": 62
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 996.7395935058594,
      "epoch": 0.09448818897637795,
      "grad_norm": 0.7355348323994314,
      "kl": 0.30078125,
      "learning_rate": 1.8805970149253735e-05,
      "loss": 0.0326,
      "reward": 1.057679921388626,
      "reward_std": 0.15489091351628304,
      "rewards/accuracy_reward": 0.0052083334885537624,
      "rewards/reasoning_steps_reward": 0.9878472238779068,
      "rewards/repetition_penalty_reward": -0.21271947026252747,
      "rewards/tag_count_reward": 0.2773437574505806,
      "step": 63
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 917.3229370117188,
      "epoch": 0.09598800149981253,
      "grad_norm": 0.5135310040841573,
      "kl": 0.3583984375,
      "learning_rate": 1.9104477611940298e-05,
      "loss": 0.0146,
      "reward": 1.2951524555683136,
      "reward_std": 0.2529895007610321,
      "rewards/accuracy_reward": 0.010416666977107525,
      "rewards/reasoning_steps_reward": 0.9340278059244156,
      "rewards/repetition_penalty_reward": -0.051635801792144775,
      "rewards/tag_count_reward": 0.4023437574505806,
      "step": 64
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 826.890625,
      "epoch": 0.09748781402324709,
      "grad_norm": 336.6953521537797,
      "kl": 19.52685546875,
      "learning_rate": 1.9402985074626868e-05,
      "loss": 0.7842,
      "reward": 1.3793235123157501,
      "reward_std": 0.32739875465631485,
      "rewards/accuracy_reward": 0.04687500139698386,
      "rewards/reasoning_steps_reward": 0.852430522441864,
      "rewards/repetition_penalty_reward": -0.02258623158559203,
      "rewards/tag_count_reward": 0.5026041865348816,
      "step": 65
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 658.5781402587891,
      "epoch": 0.09898762654668167,
      "grad_norm": 6.106229371377812,
      "kl": 0.70361328125,
      "learning_rate": 1.9701492537313435e-05,
      "loss": 0.1711,
      "reward": 1.5613079369068146,
      "reward_std": 0.4881432354450226,
      "rewards/accuracy_reward": 0.1822916716337204,
      "rewards/reasoning_steps_reward": 0.7309027910232544,
      "rewards/repetition_penalty_reward": -0.013344902312383056,
      "rewards/tag_count_reward": 0.6614583432674408,
      "step": 66
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 599.6562805175781,
      "epoch": 0.10048743907011623,
      "grad_norm": 3.5990895260846902,
      "kl": 0.43603515625,
      "learning_rate": 2e-05,
      "loss": 0.0834,
      "reward": 1.4973010122776031,
      "reward_std": 0.4555082842707634,
      "rewards/accuracy_reward": 0.1718750074505806,
      "rewards/reasoning_steps_reward": 0.630208358168602,
      "rewards/repetition_penalty_reward": -0.01051152846775949,
      "rewards/tag_count_reward": 0.7057291865348816,
      "step": 67
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 480.21876525878906,
      "epoch": 0.1019872515935508,
      "grad_norm": 1.317667285117624,
      "kl": 0.41748046875,
      "learning_rate": 1.9999862464405377e-05,
      "loss": 0.1277,
      "reward": 1.6170935332775116,
      "reward_std": 0.5107096880674362,
      "rewards/accuracy_reward": 0.276041679084301,
      "rewards/reasoning_steps_reward": 0.5694444477558136,
      "rewards/repetition_penalty_reward": -0.013548914226703346,
      "rewards/tag_count_reward": 0.78515625,
      "step": 68
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 345.32813262939453,
      "epoch": 0.10348706411698538,
      "grad_norm": 1.7734343075276982,
      "kl": 0.39208984375,
      "learning_rate": 1.9999449861404716e-05,
      "loss": 0.1866,
      "reward": 1.814243733882904,
      "reward_std": 0.5948401093482971,
      "rewards/accuracy_reward": 0.354166679084301,
      "rewards/reasoning_steps_reward": 0.6111111044883728,
      "rewards/repetition_penalty_reward": -0.01691950182430446,
      "rewards/tag_count_reward": 0.8658854365348816,
      "step": 69
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 346.2291717529297,
      "epoch": 0.10498687664041995,
      "grad_norm": 3.2645219448529943,
      "kl": 0.3359375,
      "learning_rate": 1.999876220234753e-05,
      "loss": 0.3835,
      "reward": 2.1113045811653137,
      "reward_std": 0.6916591078042984,
      "rewards/accuracy_reward": 0.5312500074505806,
      "rewards/reasoning_steps_reward": 0.7447917312383652,
      "rewards/repetition_penalty_reward": -0.02281005564145744,
      "rewards/tag_count_reward": 0.8580729514360428,
      "step": 70
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 396.17708587646484,
      "epoch": 0.10648668916385452,
      "grad_norm": 3.618615404780119,
      "kl": 0.33447265625,
      "learning_rate": 1.999779950614934e-05,
      "loss": 0.4845,
      "reward": 2.033944606781006,
      "reward_std": 0.7569809406995773,
      "rewards/accuracy_reward": 0.4843750149011612,
      "rewards/reasoning_steps_reward": 0.763888880610466,
      "rewards/repetition_penalty_reward": -0.02421517251059413,
      "rewards/tag_count_reward": 0.809895858168602,
      "step": 71
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 469.48438262939453,
      "epoch": 0.10798650168728909,
      "grad_norm": 296.41565415069374,
      "kl": 6.908203125,
      "learning_rate": 1.999656179929115e-05,
      "loss": 0.9572,
      "reward": 1.6951223015785217,
      "reward_std": 0.7641638964414597,
      "rewards/accuracy_reward": 0.2500000111758709,
      "rewards/reasoning_steps_reward": 0.8385416716337204,
      "rewards/repetition_penalty_reward": -0.017117327079176903,
      "rewards/tag_count_reward": 0.6236979365348816,
      "step": 72
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 533.2343978881836,
      "epoch": 0.10948631421072366,
      "grad_norm": 2.6480944796648282,
      "kl": 0.9912109375,
      "learning_rate": 1.9995049115818706e-05,
      "loss": 0.3052,
      "reward": 1.2268942147493362,
      "reward_std": 0.9322899430990219,
      "rewards/accuracy_reward": 0.20312500558793545,
      "rewards/reasoning_steps_reward": 0.6597222536802292,
      "rewards/repetition_penalty_reward": -0.013557222904637456,
      "rewards/tag_count_reward": 0.3776041716337204,
      "step": 73
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 329.98439025878906,
      "epoch": 0.11098612673415822,
      "grad_norm": 399.79727881978687,
      "kl": 12.41015625,
      "learning_rate": 1.9993261497341575e-05,
      "loss": 1.299,
      "reward": 0.8436856269836426,
      "reward_std": 0.9532901048660278,
      "rewards/accuracy_reward": 0.1406250037252903,
      "rewards/reasoning_steps_reward": 0.49305559694767,
      "rewards/repetition_penalty_reward": -0.0087449811398983,
      "rewards/tag_count_reward": 0.2187500037252903,
      "step": 74
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 197.5885467529297,
      "epoch": 0.1124859392575928,
      "grad_norm": 11.601181346751435,
      "kl": 4.046875,
      "learning_rate": 1.9991198993031992e-05,
      "loss": 0.1052,
      "reward": 0.5184430181980133,
      "reward_std": 0.8223861902952194,
      "rewards/accuracy_reward": 0.08333333535119891,
      "rewards/reasoning_steps_reward": 0.3211805745959282,
      "rewards/repetition_penalty_reward": -0.005862575490027666,
      "rewards/tag_count_reward": 0.11979166977107525,
      "step": 75
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 169.49480056762695,
      "epoch": 0.11398575178102738,
      "grad_norm": 4.174677375261219,
      "kl": 2.41796875,
      "learning_rate": 1.99888616596235e-05,
      "loss": -0.0061,
      "reward": 0.7175929397344589,
      "reward_std": 1.0486825108528137,
      "rewards/accuracy_reward": 0.14583334140479565,
      "rewards/reasoning_steps_reward": 0.3246527947485447,
      "rewards/repetition_penalty_reward": -0.008101542014628649,
      "rewards/tag_count_reward": 0.2552083395421505,
      "step": 76
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 175.0677146911621,
      "epoch": 0.11548556430446194,
      "grad_norm": 3.753658615227732,
      "kl": 2.3984375,
      "learning_rate": 1.9986249561409415e-05,
      "loss": -0.0758,
      "reward": 0.7132928371429443,
      "reward_std": 0.9969339072704315,
      "rewards/accuracy_reward": 0.1614583395421505,
      "rewards/reasoning_steps_reward": 0.3628472313284874,
      "rewards/repetition_penalty_reward": -0.012835650937631726,
      "rewards/tag_count_reward": 0.2018229216337204,
      "step": 77
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 189.9114646911621,
      "epoch": 0.11698537682789652,
      "grad_norm": 31.158238324524778,
      "kl": 4.71484375,
      "learning_rate": 1.998336277024103e-05,
      "loss": -0.0148,
      "reward": 1.2587463855743408,
      "reward_std": 1.1734082400798798,
      "rewards/accuracy_reward": 0.3125000074505806,
      "rewards/reasoning_steps_reward": 0.5677083730697632,
      "rewards/repetition_penalty_reward": -0.029014051891863346,
      "rewards/tag_count_reward": 0.4075521007180214,
      "step": 78
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 226.31250381469727,
      "epoch": 0.11848518935133108,
      "grad_norm": 59.19488601013592,
      "kl": 7.015625,
      "learning_rate": 1.998020136552566e-05,
      "loss": 0.0999,
      "reward": 1.4918034672737122,
      "reward_std": 1.0446466207504272,
      "rewards/accuracy_reward": 0.3333333432674408,
      "rewards/reasoning_steps_reward": 0.7187500298023224,
      "rewards/repetition_penalty_reward": -0.07199861854314804,
      "rewards/tag_count_reward": 0.5117187649011612,
      "step": 79
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 298.35939025878906,
      "epoch": 0.11998500187476566,
      "grad_norm": 77.22679135587332,
      "kl": 1.4228515625,
      "learning_rate": 1.9976765434224426e-05,
      "loss": 0.1166,
      "reward": 2.0578636527061462,
      "reward_std": 1.055399090051651,
      "rewards/accuracy_reward": 0.583333358168602,
      "rewards/reasoning_steps_reward": 0.8281250298023224,
      "rewards/repetition_penalty_reward": -0.06974058039486408,
      "rewards/tag_count_reward": 0.716145858168602,
      "step": 80
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 343.8645935058594,
      "epoch": 0.12148481439820022,
      "grad_norm": 492.7397963149802,
      "kl": 4.455078125,
      "learning_rate": 1.9973055070849912e-05,
      "loss": 0.5194,
      "reward": 2.197775959968567,
      "reward_std": 0.8521990180015564,
      "rewards/accuracy_reward": 0.59375,
      "rewards/reasoning_steps_reward": 0.913194477558136,
      "rewards/repetition_penalty_reward": -0.12166857533156872,
      "rewards/tag_count_reward": 0.8125000298023224,
      "step": 81
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 370.52083587646484,
      "epoch": 0.1229846269216348,
      "grad_norm": 4.669423434787285,
      "kl": 1.5625,
      "learning_rate": 1.996907037746352e-05,
      "loss": 0.2641,
      "reward": 2.129128336906433,
      "reward_std": 0.7840546369552612,
      "rewards/accuracy_reward": 0.5000000074505806,
      "rewards/reasoning_steps_reward": 0.9461805522441864,
      "rewards/repetition_penalty_reward": -0.14647945389151573,
      "rewards/tag_count_reward": 0.8294270932674408,
      "step": 82
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 353.7447967529297,
      "epoch": 0.12448443944506937,
      "grad_norm": 20.98608213741596,
      "kl": 0.47705078125,
      "learning_rate": 1.9964811463672685e-05,
      "loss": 0.2863,
      "reward": 2.022731065750122,
      "reward_std": 0.637829914689064,
      "rewards/accuracy_reward": 0.3645833432674408,
      "rewards/reasoning_steps_reward": 0.9583333432674408,
      "rewards/repetition_penalty_reward": -0.16476891934871674,
      "rewards/tag_count_reward": 0.864583358168602,
      "step": 83
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 330.2395935058594,
      "epoch": 0.12598425196850394,
      "grad_norm": 5.548707042279948,
      "kl": 0.576171875,
      "learning_rate": 1.996027844662785e-05,
      "loss": 0.3677,
      "reward": 2.218412935733795,
      "reward_std": 0.7215779423713684,
      "rewards/accuracy_reward": 0.557291679084301,
      "rewards/reasoning_steps_reward": 0.9548611491918564,
      "rewards/repetition_penalty_reward": -0.17525040730834007,
      "rewards/tag_count_reward": 0.8815104365348816,
      "step": 84
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 347.31251525878906,
      "epoch": 0.1274840644919385,
      "grad_norm": 35.58826029321981,
      "kl": 1.0712890625,
      "learning_rate": 1.9955471451019264e-05,
      "loss": 0.4448,
      "reward": 2.138911247253418,
      "reward_std": 0.6522012650966644,
      "rewards/accuracy_reward": 0.4687500074505806,
      "rewards/reasoning_steps_reward": 0.9357638955116272,
      "rewards/repetition_penalty_reward": -0.1510193683207035,
      "rewards/tag_count_reward": 0.8854166716337204,
      "step": 85
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 331.0833435058594,
      "epoch": 0.1289838770153731,
      "grad_norm": 10.955955280940925,
      "kl": 1.5302734375,
      "learning_rate": 1.995039060907352e-05,
      "loss": 0.4797,
      "reward": 2.054484724998474,
      "reward_std": 0.8432440906763077,
      "rewards/accuracy_reward": 0.463541679084301,
      "rewards/reasoning_steps_reward": 0.8645833730697632,
      "rewards/repetition_penalty_reward": -0.10436956211924553,
      "rewards/tag_count_reward": 0.8307292014360428,
      "step": 86
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 310.32813262939453,
      "epoch": 0.13048368953880765,
      "grad_norm": 64784.39467186327,
      "kl": 333.48828125,
      "learning_rate": 1.994503606054994e-05,
      "loss": 27.6513,
      "reward": 2.2711276412010193,
      "reward_std": 0.7795832455158234,
      "rewards/accuracy_reward": 0.5885416865348816,
      "rewards/reasoning_steps_reward": 0.9166666716337204,
      "rewards/repetition_penalty_reward": -0.11559123173356056,
      "rewards/tag_count_reward": 0.8815104514360428,
      "step": 87
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 311.98438262939453,
      "epoch": 0.13198350206224221,
      "grad_norm": 36649.02893473602,
      "kl": 143.029296875,
      "learning_rate": 1.9939407952736737e-05,
      "loss": 15.8982,
      "reward": 1.9791134297847748,
      "reward_std": 0.6517055183649063,
      "rewards/accuracy_reward": 0.3489583432674408,
      "rewards/reasoning_steps_reward": 0.9045139402151108,
      "rewards/repetition_penalty_reward": -0.13503596186637878,
      "rewards/tag_count_reward": 0.860677108168602,
      "step": 88
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 329.5052185058594,
      "epoch": 0.13348331458567678,
      "grad_norm": 892.1216234188144,
      "kl": 7.07421875,
      "learning_rate": 1.9933506440446932e-05,
      "loss": 0.984,
      "reward": 1.9748746156692505,
      "reward_std": 0.6786476969718933,
      "rewards/accuracy_reward": 0.3541666716337204,
      "rewards/reasoning_steps_reward": 0.9079861491918564,
      "rewards/repetition_penalty_reward": -0.11670541763305664,
      "rewards/tag_count_reward": 0.829427108168602,
      "step": 89
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 291.7604293823242,
      "epoch": 0.13498312710911137,
      "grad_norm": 49.617196414779116,
      "kl": 2.85791015625,
      "learning_rate": 1.992733168601413e-05,
      "loss": 0.3242,
      "reward": 1.9374914765357971,
      "reward_std": 0.5130168125033379,
      "rewards/accuracy_reward": 0.27604167349636555,
      "rewards/reasoning_steps_reward": 0.954861119389534,
      "rewards/repetition_penalty_reward": -0.1592968888580799,
      "rewards/tag_count_reward": 0.8658854365348816,
      "step": 90
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 293.4739685058594,
      "epoch": 0.13648293963254593,
      "grad_norm": 11.069036826310084,
      "kl": 0.42626953125,
      "learning_rate": 1.9920883859288035e-05,
      "loss": 0.3009,
      "reward": 1.941705048084259,
      "reward_std": 0.5944435596466064,
      "rewards/accuracy_reward": 0.3177083395421505,
      "rewards/reasoning_steps_reward": 0.9322917014360428,
      "rewards/repetition_penalty_reward": -0.14032622054219246,
      "rewards/tag_count_reward": 0.8320312649011612,
      "step": 91
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 302.9948043823242,
      "epoch": 0.1379827521559805,
      "grad_norm": 70.67912729974059,
      "kl": 2.607421875,
      "learning_rate": 1.991416313762978e-05,
      "loss": 0.3458,
      "reward": 1.9879169762134552,
      "reward_std": 0.5511599630117416,
      "rewards/accuracy_reward": 0.3385416753590107,
      "rewards/reasoning_steps_reward": 0.9270833134651184,
      "rewards/repetition_penalty_reward": -0.14489559456706047,
      "rewards/tag_count_reward": 0.8671875149011612,
      "step": 92
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 327.0052185058594,
      "epoch": 0.13948256467941508,
      "grad_norm": 18.812219164179464,
      "kl": 1.408203125,
      "learning_rate": 1.990716970590706e-05,
      "loss": 0.3788,
      "reward": 2.0359848737716675,
      "reward_std": 0.5520138740539551,
      "rewards/accuracy_reward": 0.36979167675599456,
      "rewards/reasoning_steps_reward": 0.9322917461395264,
      "rewards/repetition_penalty_reward": -0.1345881223678589,
      "rewards/tag_count_reward": 0.8684896230697632,
      "step": 93
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 359.4010543823242,
      "epoch": 0.14098237720284965,
      "grad_norm": 14.184878742271176,
      "kl": 0.765625,
      "learning_rate": 1.989990375648903e-05,
      "loss": 0.324,
      "reward": 2.120865046977997,
      "reward_std": 0.6140560433268547,
      "rewards/accuracy_reward": 0.427083358168602,
      "rewards/reasoning_steps_reward": 0.9531250298023224,
      "rewards/repetition_penalty_reward": -0.12653085589408875,
      "rewards/tag_count_reward": 0.8671875149011612,
      "step": 94
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 372.76564025878906,
      "epoch": 0.1424821897262842,
      "grad_norm": 33.74529208746235,
      "kl": 5.03515625,
      "learning_rate": 1.9892365489241023e-05,
      "loss": 0.4259,
      "reward": 2.2041754722595215,
      "reward_std": 0.6197474002838135,
      "rewards/accuracy_reward": 0.5260416939854622,
      "rewards/reasoning_steps_reward": 0.9583333730697632,
      "rewards/repetition_penalty_reward": -0.1499912552535534,
      "rewards/tag_count_reward": 0.8697916865348816,
      "step": 95
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 409.7604293823242,
      "epoch": 0.1439820022497188,
      "grad_norm": 60.09340508863677,
      "kl": 0.78662109375,
      "learning_rate": 1.988455511151906e-05,
      "loss": 0.1251,
      "reward": 2.28587007522583,
      "reward_std": 0.5608475357294083,
      "rewards/accuracy_reward": 0.557291679084301,
      "rewards/reasoning_steps_reward": 0.9600694924592972,
      "rewards/repetition_penalty_reward": -0.1312306523323059,
      "rewards/tag_count_reward": 0.8997395932674408,
      "step": 96
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 485.81250762939453,
      "epoch": 0.14548181477315336,
      "grad_norm": 39.25906795250181,
      "kl": 0.7392578125,
      "learning_rate": 1.987647283816412e-05,
      "loss": 0.2008,
      "reward": 2.1415184140205383,
      "reward_std": 0.593282975256443,
      "rewards/accuracy_reward": 0.4270833432674408,
      "rewards/reasoning_steps_reward": 0.947916716337204,
      "rewards/repetition_penalty_reward": -0.1084816437214613,
      "rewards/tag_count_reward": 0.8750000298023224,
      "step": 97
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 527.7396087646484,
      "epoch": 0.14698162729658792,
      "grad_norm": 23.53154353337868,
      "kl": 1.400390625,
      "learning_rate": 1.9868118891496268e-05,
      "loss": 0.2105,
      "reward": 1.9524771869182587,
      "reward_std": 0.5782586634159088,
      "rewards/accuracy_reward": 0.2760416716337204,
      "rewards/reasoning_steps_reward": 0.9444444328546524,
      "rewards/repetition_penalty_reward": -0.11045688763260841,
      "rewards/tag_count_reward": 0.8424479514360428,
      "step": 98
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 567.171875,
      "epoch": 0.1484814398200225,
      "grad_norm": 32.44194296184474,
      "kl": 3.03515625,
      "learning_rate": 1.98594935013085e-05,
      "loss": 0.3744,
      "reward": 2.2206438183784485,
      "reward_std": 0.6493653506040573,
      "rewards/accuracy_reward": 0.520833358168602,
      "rewards/reasoning_steps_reward": 0.9670139253139496,
      "rewards/repetition_penalty_reward": -0.1200681421905756,
      "rewards/tag_count_reward": 0.852864608168602,
      "step": 99
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 613.2604522705078,
      "epoch": 0.14998125234345708,
      "grad_norm": 10134.408469401435,
      "kl": 115.4140625,
      "learning_rate": 1.985059690486045e-05,
      "loss": 6.0425,
      "reward": 1.976808786392212,
      "reward_std": 0.5220333635807037,
      "rewards/accuracy_reward": 0.2604166716337204,
      "rewards/reasoning_steps_reward": 0.9670139253139496,
      "rewards/repetition_penalty_reward": -0.11390306055545807,
      "rewards/tag_count_reward": 0.8632812649011612,
      "step": 100
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 638.0677337646484,
      "epoch": 0.15148106486689164,
      "grad_norm": 11770.662483971002,
      "kl": 115.66796875,
      "learning_rate": 1.9841429346871863e-05,
      "loss": 7.1104,
      "reward": 2.012724369764328,
      "reward_std": 0.6818199008703232,
      "rewards/accuracy_reward": 0.3750000037252903,
      "rewards/reasoning_steps_reward": 0.9218750447034836,
      "rewards/repetition_penalty_reward": -0.10706735588610172,
      "rewards/tag_count_reward": 0.8229166865348816,
      "step": 101
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 660.0156402587891,
      "epoch": 0.1529808773903262,
      "grad_norm": 2359.6075026833482,
      "kl": 29.76953125,
      "learning_rate": 1.9831991079515836e-05,
      "loss": 1.7284,
      "reward": 2.195272386074066,
      "reward_std": 0.7458173632621765,
      "rewards/accuracy_reward": 0.5416666939854622,
      "rewards/reasoning_steps_reward": 0.9340277910232544,
      "rewards/repetition_penalty_reward": -0.11635957658290863,
      "rewards/tag_count_reward": 0.8359375298023224,
      "step": 102
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 690.4166870117188,
      "epoch": 0.1544806899137608,
      "grad_norm": 34.92066839587673,
      "kl": 3.017578125,
      "learning_rate": 1.982228236241192e-05,
      "loss": 0.2735,
      "reward": 2.1599226593971252,
      "reward_std": 0.7309878766536713,
      "rewards/accuracy_reward": 0.479166679084301,
      "rewards/reasoning_steps_reward": 0.942708358168602,
      "rewards/repetition_penalty_reward": -0.11742118000984192,
      "rewards/tag_count_reward": 0.8554687798023224,
      "step": 103
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 669.7448120117188,
      "epoch": 0.15598050243719536,
      "grad_norm": 586.5134421102061,
      "kl": 9.271484375,
      "learning_rate": 1.9812303462618945e-05,
      "loss": 0.5437,
      "reward": 1.991296112537384,
      "reward_std": 0.5067874565720558,
      "rewards/accuracy_reward": 0.2916666818782687,
      "rewards/reasoning_steps_reward": 0.9652778208255768,
      "rewards/repetition_penalty_reward": -0.11460676416754723,
      "rewards/tag_count_reward": 0.848958358168602,
      "step": 104
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 727.6771087646484,
      "epoch": 0.15748031496062992,
      "grad_norm": 398.23420182364833,
      "kl": 10.8125,
      "learning_rate": 1.9802054654627694e-05,
      "loss": 0.5388,
      "reward": 1.8793240189552307,
      "reward_std": 0.7508310377597809,
      "rewards/accuracy_reward": 0.3020833432674408,
      "rewards/reasoning_steps_reward": 0.9149306267499924,
      "rewards/repetition_penalty_reward": -0.0942003782838583,
      "rewards/tag_count_reward": 0.7565104365348816,
      "step": 105
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 729.7760620117188,
      "epoch": 0.15898012748406448,
      "grad_norm": 14.580781126168798,
      "kl": 0.92138671875,
      "learning_rate": 1.9791536220353355e-05,
      "loss": 0.1662,
      "reward": 1.9787148237228394,
      "reward_std": 0.6732726097106934,
      "rewards/accuracy_reward": 0.3437500149011612,
      "rewards/reasoning_steps_reward": 0.9531250149011612,
      "rewards/repetition_penalty_reward": -0.10201445035636425,
      "rewards/tag_count_reward": 0.7838541865348816,
      "step": 106
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 700.3698120117188,
      "epoch": 0.16047994000749907,
      "grad_norm": 78.39112156030271,
      "kl": 4.37890625,
      "learning_rate": 1.9780748449127745e-05,
      "loss": 0.3547,
      "reward": 1.8721507489681244,
      "reward_std": 0.7452912926673889,
      "rewards/accuracy_reward": 0.3072916753590107,
      "rewards/reasoning_steps_reward": 0.9253472685813904,
      "rewards/repetition_penalty_reward": -0.08184238523244858,
      "rewards/tag_count_reward": 0.7213541865348816,
      "step": 107
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 751.7448120117188,
      "epoch": 0.16197975253093364,
      "grad_norm": 23.15048932086659,
      "kl": 3.77734375,
      "learning_rate": 1.976969163769137e-05,
      "loss": 0.3573,
      "reward": 1.7692583501338959,
      "reward_std": 0.6871030032634735,
      "rewards/accuracy_reward": 0.2135416753590107,
      "rewards/reasoning_steps_reward": 0.9375000447034836,
      "rewards/repetition_penalty_reward": -0.0770958885550499,
      "rewards/tag_count_reward": 0.6953125149011612,
      "step": 108
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 666.0625152587891,
      "epoch": 0.1634795650543682,
      "grad_norm": 13.438163054952135,
      "kl": 0.791015625,
      "learning_rate": 1.9758366090185255e-05,
      "loss": 0.194,
      "reward": 2.101614534854889,
      "reward_std": 0.779254287481308,
      "rewards/accuracy_reward": 0.4739583432674408,
      "rewards/reasoning_steps_reward": 0.9357639253139496,
      "rewards/repetition_penalty_reward": -0.08805570006370544,
      "rewards/tag_count_reward": 0.7799479365348816,
      "step": 109
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 657.4218902587891,
      "epoch": 0.1649793775778028,
      "grad_norm": 19.486727225701877,
      "kl": 1.15625,
      "learning_rate": 1.974677211814259e-05,
      "loss": 0.2437,
      "reward": 2.0427930653095245,
      "reward_std": 0.7101475596427917,
      "rewards/accuracy_reward": 0.4062500149011612,
      "rewards/reasoning_steps_reward": 0.9270833879709244,
      "rewards/repetition_penalty_reward": -0.11345694959163666,
      "rewards/tag_count_reward": 0.8229166865348816,
      "step": 110
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 616.7968902587891,
      "epoch": 0.16647919010123735,
      "grad_norm": 43.9361877854642,
      "kl": 5.32421875,
      "learning_rate": 1.973491004048014e-05,
      "loss": 0.3651,
      "reward": 2.290136158466339,
      "reward_std": 0.5547062531113625,
      "rewards/accuracy_reward": 0.5416666716337204,
      "rewards/reasoning_steps_reward": 0.9531249850988388,
      "rewards/repetition_penalty_reward": -0.11481184512376785,
      "rewards/tag_count_reward": 0.9101562798023224,
      "step": 111
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 616.4948120117188,
      "epoch": 0.1679790026246719,
      "grad_norm": 25.513657082760023,
      "kl": 2.0654296875,
      "learning_rate": 1.9722780183489477e-05,
      "loss": 0.1826,
      "reward": 2.2123002409934998,
      "reward_std": 0.6505413353443146,
      "rewards/accuracy_reward": 0.4687500149011612,
      "rewards/reasoning_steps_reward": 0.953125,
      "rewards/repetition_penalty_reward": -0.13665815442800522,
      "rewards/tag_count_reward": 0.927083358168602,
      "step": 112
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 639.6198120117188,
      "epoch": 0.16947881514810648,
      "grad_norm": 4.0215101485261595,
      "kl": 1.24365234375,
      "learning_rate": 1.9710382880828028e-05,
      "loss": 0.1482,
      "reward": 2.214249849319458,
      "reward_std": 0.6732968837022781,
      "rewards/accuracy_reward": 0.5000000074505806,
      "rewards/reasoning_steps_reward": 0.9461805522441864,
      "rewards/repetition_penalty_reward": -0.1433890499174595,
      "rewards/tag_count_reward": 0.9114583432674408,
      "step": 113
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 692.9062652587891,
      "epoch": 0.17097862767154107,
      "grad_norm": 78.14938942373506,
      "kl": 5.59765625,
      "learning_rate": 1.969771847350987e-05,
      "loss": 0.4814,
      "reward": 2.0078442692756653,
      "reward_std": 0.8100399523973465,
      "rewards/accuracy_reward": 0.4218750149011612,
      "rewards/reasoning_steps_reward": 0.9253472238779068,
      "rewards/repetition_penalty_reward": -0.0971906129270792,
      "rewards/tag_count_reward": 0.7578125149011612,
      "step": 114
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 873.9166870117188,
      "epoch": 0.17247844019497563,
      "grad_norm": 390.6492744892592,
      "kl": 20.15625,
      "learning_rate": 1.968478730989636e-05,
      "loss": 0.931,
      "reward": 1.3626113533973694,
      "reward_std": 0.49340808391571045,
      "rewards/accuracy_reward": 0.07291666930541396,
      "rewards/reasoning_steps_reward": 0.9409722238779068,
      "rewards/repetition_penalty_reward": -0.03799628745764494,
      "rewards/tag_count_reward": 0.3867187574505806,
      "step": 115
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 889.4166870117188,
      "epoch": 0.1739782527184102,
      "grad_norm": 46.43260724644146,
      "kl": 7.609375,
      "learning_rate": 1.9671589745686563e-05,
      "loss": 0.3225,
      "reward": 1.1709461510181427,
      "reward_std": 0.3466350585222244,
      "rewards/accuracy_reward": 0.010416666977107525,
      "rewards/reasoning_steps_reward": 0.9201389402151108,
      "rewards/repetition_penalty_reward": -0.023932393174618483,
      "rewards/tag_count_reward": 0.2643229216337204,
      "step": 116
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 855.1302185058594,
      "epoch": 0.17547806524184478,
      "grad_norm": 19.881821438298648,
      "kl": 2.70703125,
      "learning_rate": 1.965812614390743e-05,
      "loss": 0.1427,
      "reward": 1.1638481318950653,
      "reward_std": 0.343635730445385,
      "rewards/accuracy_reward": 0.010416666977107525,
      "rewards/reasoning_steps_reward": 0.9027778059244156,
      "rewards/repetition_penalty_reward": -0.027992176823318005,
      "rewards/tag_count_reward": 0.2786458358168602,
      "step": 117
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 837.4010620117188,
      "epoch": 0.17697787776527935,
      "grad_norm": 10.649672323547897,
      "kl": 1.78125,
      "learning_rate": 1.9644396874903865e-05,
      "loss": 0.056,
      "reward": 1.1770220398902893,
      "reward_std": 0.30946608632802963,
      "rewards/accuracy_reward": 0.0052083334885537624,
      "rewards/reasoning_steps_reward": 0.927083358168602,
      "rewards/repetition_penalty_reward": -0.028707201592624187,
      "rewards/tag_count_reward": 0.2734375074505806,
      "step": 118
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 783.4479370117188,
      "epoch": 0.1784776902887139,
      "grad_norm": 6.976349962907348,
      "kl": 1.45703125,
      "learning_rate": 1.9630402316328506e-05,
      "loss": 0.0257,
      "reward": 1.2304543852806091,
      "reward_std": 0.30063216388225555,
      "rewards/accuracy_reward": 0.0052083334885537624,
      "rewards/reasoning_steps_reward": 0.9461805671453476,
      "rewards/repetition_penalty_reward": -0.033434574957937,
      "rewards/tag_count_reward": 0.3125000149011612,
      "step": 119
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 655.8281402587891,
      "epoch": 0.17997750281214847,
      "grad_norm": 4.798337799999206,
      "kl": 0.5478515625,
      "learning_rate": 1.9616142853131342e-05,
      "loss": 0.1748,
      "reward": 2.003317028284073,
      "reward_std": 0.682994619011879,
      "rewards/accuracy_reward": 0.3489583432674408,
      "rewards/reasoning_steps_reward": 0.9826389253139496,
      "rewards/repetition_penalty_reward": -0.1004156544804573,
      "rewards/tag_count_reward": 0.7721354514360428,
      "step": 120
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 596.4583587646484,
      "epoch": 0.18147731533558306,
      "grad_norm": 7.178534631101688,
      "kl": 0.65283203125,
      "learning_rate": 1.9601618877549113e-05,
      "loss": 0.1098,
      "reward": 2.141425609588623,
      "reward_std": 0.6803022921085358,
      "rewards/accuracy_reward": 0.453125,
      "rewards/reasoning_steps_reward": 0.9548611640930176,
      "rewards/repetition_penalty_reward": -0.12593566998839378,
      "rewards/tag_count_reward": 0.8593750298023224,
      "step": 121
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 585.9843902587891,
      "epoch": 0.18297712785901762,
      "grad_norm": 5.756866189784068,
      "kl": 1.4296875,
      "learning_rate": 1.9586830789094548e-05,
      "loss": 0.0643,
      "reward": 2.145204782485962,
      "reward_std": 0.6050543040037155,
      "rewards/accuracy_reward": 0.4322916939854622,
      "rewards/reasoning_steps_reward": 0.9687500298023224,
      "rewards/repetition_penalty_reward": -0.13474315032362938,
      "rewards/tag_count_reward": 0.8789062649011612,
      "step": 122
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 534.2760543823242,
      "epoch": 0.1844769403824522,
      "grad_norm": 104.47967057005842,
      "kl": 4.3759765625,
      "learning_rate": 1.9571778994545356e-05,
      "loss": 0.2084,
      "reward": 2.2705207467079163,
      "reward_std": 0.7510545551776886,
      "rewards/accuracy_reward": 0.5729166865348816,
      "rewards/reasoning_steps_reward": 0.9409722834825516,
      "rewards/repetition_penalty_reward": -0.11706613004207611,
      "rewards/tag_count_reward": 0.8736979365348816,
      "step": 123
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 534.5937652587891,
      "epoch": 0.18597675290588678,
      "grad_norm": 14.350020581816006,
      "kl": 1.861328125,
      "learning_rate": 1.9556463907933038e-05,
      "loss": 0.0927,
      "reward": 2.2636232376098633,
      "reward_std": 0.6755192577838898,
      "rewards/accuracy_reward": 0.520833358168602,
      "rewards/reasoning_steps_reward": 0.9600694328546524,
      "rewards/repetition_penalty_reward": -0.12092556245625019,
      "rewards/tag_count_reward": 0.9036458730697632,
      "step": 124
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 565.2968902587891,
      "epoch": 0.18747656542932134,
      "grad_norm": 3.30622411421042,
      "kl": 1.943359375,
      "learning_rate": 1.9540885950531507e-05,
      "loss": 0.1088,
      "reward": 2.27703320980072,
      "reward_std": 0.6729772835969925,
      "rewards/accuracy_reward": 0.5677083432674408,
      "rewards/reasoning_steps_reward": 0.9392361342906952,
      "rewards/repetition_penalty_reward": -0.13225507363677025,
      "rewards/tag_count_reward": 0.9023437649011612,
      "step": 125
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 582.3177185058594,
      "epoch": 0.1889763779527559,
      "grad_norm": 13.34144394107955,
      "kl": 40.078125,
      "learning_rate": 1.9525045550845482e-05,
      "loss": 0.1878,
      "reward": 2.363175332546234,
      "reward_std": 0.654795840382576,
      "rewards/accuracy_reward": 0.6510416716337204,
      "rewards/reasoning_steps_reward": 0.9409722536802292,
      "rewards/repetition_penalty_reward": -0.15722395852208138,
      "rewards/tag_count_reward": 0.9283854514360428,
      "step": 126
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 574.8698120117188,
      "epoch": 0.19047619047619047,
      "grad_norm": 21.137483639050657,
      "kl": 5.4140625,
      "learning_rate": 1.9508943144598726e-05,
      "loss": 0.2436,
      "reward": 2.1040448546409607,
      "reward_std": 0.7035096734762192,
      "rewards/accuracy_reward": 0.4062500149011612,
      "rewards/reasoning_steps_reward": 0.9270834028720856,
      "rewards/repetition_penalty_reward": -0.13814280182123184,
      "rewards/tag_count_reward": 0.9088541865348816,
      "step": 127
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 531.4687652587891,
      "epoch": 0.19197600299962506,
      "grad_norm": 4.455726874489183,
      "kl": 0.81103515625,
      "learning_rate": 1.9492579174722043e-05,
      "loss": -0.0353,
      "reward": 2.3104677200317383,
      "reward_std": 0.5839991569519043,
      "rewards/accuracy_reward": 0.5885416865348816,
      "rewards/reasoning_steps_reward": 0.9496528059244156,
      "rewards/repetition_penalty_reward": -0.16132058203220367,
      "rewards/tag_count_reward": 0.9335937798023224,
      "step": 128
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 518.4531402587891,
      "epoch": 0.19347581552305962,
      "grad_norm": 681.8928268072538,
      "kl": 14.496826171875,
      "learning_rate": 1.9475954091341098e-05,
      "loss": 0.7073,
      "reward": 2.2904911935329437,
      "reward_std": 0.6285631433129311,
      "rewards/accuracy_reward": 0.5885416865348816,
      "rewards/reasoning_steps_reward": 0.9236111342906952,
      "rewards/repetition_penalty_reward": -0.14353665709495544,
      "rewards/tag_count_reward": 0.9218750149011612,
      "step": 129
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 495.0937728881836,
      "epoch": 0.19497562804649418,
      "grad_norm": 26.634074769249203,
      "kl": 14.8203125,
      "learning_rate": 1.9459068351764032e-05,
      "loss": 0.1622,
      "reward": 2.11430823802948,
      "reward_std": 0.6787235736846924,
      "rewards/accuracy_reward": 0.4322916865348816,
      "rewards/reasoning_steps_reward": 0.923611119389534,
      "rewards/repetition_penalty_reward": -0.14263629540801048,
      "rewards/tag_count_reward": 0.9010416865348816,
      "step": 130
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 465.46876525878906,
      "epoch": 0.19647544056992877,
      "grad_norm": 3.2870541929839407,
      "kl": 1.111328125,
      "learning_rate": 1.94419224204689e-05,
      "loss": -0.1421,
      "reward": 2.2763773798942566,
      "reward_std": 0.7801897525787354,
      "rewards/accuracy_reward": 0.630208358168602,
      "rewards/reasoning_steps_reward": 0.9131944626569748,
      "rewards/repetition_penalty_reward": -0.14202555641531944,
      "rewards/tag_count_reward": 0.8750000149011612,
      "step": 131
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 467.06250762939453,
      "epoch": 0.19797525309336333,
      "grad_norm": 9.180164568388207,
      "kl": 1.109375,
      "learning_rate": 1.9424516769090863e-05,
      "loss": -0.0418,
      "reward": 1.8292181193828583,
      "reward_std": 1.029387205839157,
      "rewards/accuracy_reward": 0.4114583358168602,
      "rewards/reasoning_steps_reward": 0.82118059694767,
      "rewards/repetition_penalty_reward": -0.1039416529238224,
      "rewards/tag_count_reward": 0.700520858168602,
      "step": 132
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 455.7291793823242,
      "epoch": 0.1994750656167979,
      "grad_norm": 2.948220887698058,
      "kl": 1.56640625,
      "learning_rate": 1.9406851876409254e-05,
      "loss": -0.1012,
      "reward": 1.8498985171318054,
      "reward_std": 1.0783482491970062,
      "rewards/accuracy_reward": 0.4843750298023224,
      "rewards/reasoning_steps_reward": 0.7986111342906952,
      "rewards/repetition_penalty_reward": -0.08673347532749176,
      "rewards/tag_count_reward": 0.6536458432674408,
      "step": 133
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 456.6927261352539,
      "epoch": 0.20097487814023246,
      "grad_norm": 6.426646825304548,
      "kl": 2.6953125,
      "learning_rate": 1.938892822833437e-05,
      "loss": 0.0348,
      "reward": 2.087424159049988,
      "reward_std": 0.9264509230852127,
      "rewards/accuracy_reward": 0.578125,
      "rewards/reasoning_steps_reward": 0.8697917312383652,
      "rewards/repetition_penalty_reward": -0.11830509081482887,
      "rewards/tag_count_reward": 0.7578125149011612,
      "step": 134
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 476.8333435058594,
      "epoch": 0.20247469066366705,
      "grad_norm": 51.991305129273876,
      "kl": 22.15625,
      "learning_rate": 1.9370746317894135e-05,
      "loss": 0.6123,
      "reward": 1.7000366151332855,
      "reward_std": 0.9176836758852005,
      "rewards/accuracy_reward": 0.3072916753590107,
      "rewards/reasoning_steps_reward": 0.8246528208255768,
      "rewards/repetition_penalty_reward": -0.102480823174119,
      "rewards/tag_count_reward": 0.6705729365348816,
      "step": 135
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 570.3385543823242,
      "epoch": 0.2039745031871016,
      "grad_norm": 8.533590166150786,
      "kl": 7.75,
      "learning_rate": 1.9352306645220518e-05,
      "loss": 0.2573,
      "reward": 1.376660943031311,
      "reward_std": 0.9236660450696945,
      "rewards/accuracy_reward": 0.2187500074505806,
      "rewards/reasoning_steps_reward": 0.7638889104127884,
      "rewards/repetition_penalty_reward": -0.07603009976446629,
      "rewards/tag_count_reward": 0.4700520932674408,
      "step": 136
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 655.1042022705078,
      "epoch": 0.20547431571053618,
      "grad_norm": 4.411821253364697,
      "kl": 2.53515625,
      "learning_rate": 1.9333609717535788e-05,
      "loss": -0.0005,
      "reward": 0.9961891174316406,
      "reward_std": 0.7426871508359909,
      "rewards/accuracy_reward": 0.06250000186264515,
      "rewards/reasoning_steps_reward": 0.6788194477558136,
      "rewards/repetition_penalty_reward": -0.04070331249386072,
      "rewards/tag_count_reward": 0.295572929084301,
      "step": 137
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 592.6458435058594,
      "epoch": 0.20697412823397077,
      "grad_norm": 1.4735089906729,
      "kl": 1.529296875,
      "learning_rate": 1.931465604913856e-05,
      "loss": -0.0875,
      "reward": 1.0002163350582123,
      "reward_std": 0.7004366517066956,
      "rewards/accuracy_reward": 0.052083334885537624,
      "rewards/reasoning_steps_reward": 0.7013889253139496,
      "rewards/repetition_penalty_reward": -0.04362048767507076,
      "rewards/tag_count_reward": 0.2903645932674408,
      "step": 138
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 725.1979370117188,
      "epoch": 0.20847394075740533,
      "grad_norm": 2.9297077947117667,
      "kl": 1.375,
      "learning_rate": 1.9295446161389644e-05,
      "loss": 0.1009,
      "reward": 1.0900856852531433,
      "reward_std": 0.5820793807506561,
      "rewards/accuracy_reward": 0.05729166837409139,
      "rewards/reasoning_steps_reward": 0.8350694924592972,
      "rewards/repetition_penalty_reward": -0.07310881279408932,
      "rewards/tag_count_reward": 0.2708333395421505,
      "step": 139
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 769.4843902587891,
      "epoch": 0.2099737532808399,
      "grad_norm": 2.6937963105185894,
      "kl": 2.18359375,
      "learning_rate": 1.9275980582697707e-05,
      "loss": 0.1039,
      "reward": 1.2186349630355835,
      "reward_std": 0.5874937772750854,
      "rewards/accuracy_reward": 0.1770833395421505,
      "rewards/reasoning_steps_reward": 0.9166667014360428,
      "rewards/repetition_penalty_reward": -0.13422967679798603,
      "rewards/tag_count_reward": 0.2591145932674408,
      "step": 140
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 845.2552337646484,
      "epoch": 0.21147356580427445,
      "grad_norm": 1.622095571716092,
      "kl": 1.2060546875,
      "learning_rate": 1.9256259848504737e-05,
      "loss": 0.0388,
      "reward": 1.2769785821437836,
      "reward_std": 0.47676292806863785,
      "rewards/accuracy_reward": 0.2968750074505806,
      "rewards/reasoning_steps_reward": 0.9704861342906952,
      "rewards/repetition_penalty_reward": -0.21303880959749222,
      "rewards/tag_count_reward": 0.2226562574505806,
      "step": 141
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 826.1771087646484,
      "epoch": 0.21297337832770905,
      "grad_norm": 1.4773462554374228,
      "kl": 0.5615234375,
      "learning_rate": 1.9236284501271317e-05,
      "loss": -0.0577,
      "reward": 1.3371992409229279,
      "reward_std": 0.49016065895557404,
      "rewards/accuracy_reward": 0.494791679084301,
      "rewards/reasoning_steps_reward": 0.9843750149011612,
      "rewards/repetition_penalty_reward": -0.27738412469625473,
      "rewards/tag_count_reward": 0.1354166716337204,
      "step": 142
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 909.7500305175781,
      "epoch": 0.2144731908511436,
      "grad_norm": 0.9239369002896122,
      "kl": 0.66259765625,
      "learning_rate": 1.9216055090461693e-05,
      "loss": -0.0375,
      "reward": 1.4242282509803772,
      "reward_std": 0.5077776834368706,
      "rewards/accuracy_reward": 0.6562500149011612,
      "rewards/reasoning_steps_reward": 0.97743059694767,
      "rewards/repetition_penalty_reward": -0.32794189453125,
      "rewards/tag_count_reward": 0.11848958767950535,
      "step": 143
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 927.3541870117188,
      "epoch": 0.21597300337457817,
      "grad_norm": 0.7490847651592132,
      "kl": 0.4716796875,
      "learning_rate": 1.9195572172528678e-05,
      "loss": 0.0525,
      "reward": 1.2379167079925537,
      "reward_std": 0.5809585899114609,
      "rewards/accuracy_reward": 0.4479166716337204,
      "rewards/reasoning_steps_reward": 0.9809028059244156,
      "rewards/repetition_penalty_reward": -0.41616319864988327,
      "rewards/tag_count_reward": 0.2252604216337204,
      "step": 144
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 782.5052337646484,
      "epoch": 0.21747281589801276,
      "grad_norm": 9.579120968298097,
      "kl": 91.814208984375,
      "learning_rate": 1.9174836310898334e-05,
      "loss": 0.1581,
      "reward": 1.7532488107681274,
      "reward_std": 0.6942550539970398,
      "rewards/accuracy_reward": 0.479166679084301,
      "rewards/reasoning_steps_reward": 0.9843750149011612,
      "rewards/repetition_penalty_reward": -0.35612622648477554,
      "rewards/tag_count_reward": 0.645833358168602,
      "step": 145
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 806.2708587646484,
      "epoch": 0.21897262842144732,
      "grad_norm": 1.8428465118174489,
      "kl": 0.408447265625,
      "learning_rate": 1.9153848075954465e-05,
      "loss": 0.1999,
      "reward": 1.5887240171432495,
      "reward_std": 0.7087294459342957,
      "rewards/accuracy_reward": 0.3437500074505806,
      "rewards/reasoning_steps_reward": 0.9791667014360428,
      "rewards/repetition_penalty_reward": -0.37351563572883606,
      "rewards/tag_count_reward": 0.6393229365348816,
      "step": 146
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 754.6458435058594,
      "epoch": 0.2204724409448819,
      "grad_norm": 0.7681530290808569,
      "kl": 0.266357421875,
      "learning_rate": 1.9132608045022954e-05,
      "loss": 0.2529,
      "reward": 1.4556803405284882,
      "reward_std": 0.7271180897951126,
      "rewards/accuracy_reward": 0.3385416753590107,
      "rewards/reasoning_steps_reward": 0.987847238779068,
      "rewards/repetition_penalty_reward": -0.3368544206023216,
      "rewards/tag_count_reward": 0.4661458432674408,
      "step": 147
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 771.1458435058594,
      "epoch": 0.22197225346831645,
      "grad_norm": 1.2907787204307715,
      "kl": 0.51708984375,
      "learning_rate": 1.9111116802355853e-05,
      "loss": 0.2698,
      "reward": 0.9035700112581253,
      "reward_std": 0.7091249525547028,
      "rewards/accuracy_reward": 0.1458333395421505,
      "rewards/reasoning_steps_reward": 0.9513889253139496,
      "rewards/repetition_penalty_reward": -0.44104810059070587,
      "rewards/tag_count_reward": 0.2473958432674408,
      "step": 148
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 578.7760467529297,
      "epoch": 0.22347206599175104,
      "grad_norm": 1.002753301602421,
      "kl": 0.42529296875,
      "learning_rate": 1.9089374939115335e-05,
      "loss": 0.3394,
      "reward": 1.304781287908554,
      "reward_std": 0.7868980914354324,
      "rewards/accuracy_reward": 0.2239583395421505,
      "rewards/reasoning_steps_reward": 0.958333358168602,
      "rewards/repetition_penalty_reward": -0.3111041784286499,
      "rewards/tag_count_reward": 0.4335937649011612,
      "step": 149
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 464.1771011352539,
      "epoch": 0.2249718785151856,
      "grad_norm": 1.5602927845713264,
      "kl": 0.5703125,
      "learning_rate": 1.906738305335741e-05,
      "loss": 0.3492,
      "reward": 1.3854647278785706,
      "reward_std": 0.942527711391449,
      "rewards/accuracy_reward": 0.2656250074505806,
      "rewards/reasoning_steps_reward": 0.9010417014360428,
      "rewards/repetition_penalty_reward": -0.272087462246418,
      "rewards/tag_count_reward": 0.490885429084301,
      "step": 150
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 455.96876525878906,
      "epoch": 0.22647169103862017,
      "grad_norm": 1.4291694851849999,
      "kl": 0.609375,
      "learning_rate": 1.90451417500155e-05,
      "loss": 0.1916,
      "reward": 1.4960919618606567,
      "reward_std": 0.7673981636762619,
      "rewards/accuracy_reward": 0.2291666753590107,
      "rewards/reasoning_steps_reward": 0.8871527761220932,
      "rewards/repetition_penalty_reward": -0.24392545968294144,
      "rewards/tag_count_reward": 0.6236979365348816,
      "step": 151
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 487.4947967529297,
      "epoch": 0.22797150356205476,
      "grad_norm": 3.3680398671635974,
      "kl": 0.68359375,
      "learning_rate": 1.902265164088378e-05,
      "loss": 0.3357,
      "reward": 1.656085580587387,
      "reward_std": 0.8526331186294556,
      "rewards/accuracy_reward": 0.3333333507180214,
      "rewards/reasoning_steps_reward": 0.944444477558136,
      "rewards/repetition_penalty_reward": -0.27143188565969467,
      "rewards/tag_count_reward": 0.6497395932674408,
      "step": 152
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 541.6562652587891,
      "epoch": 0.22947131608548932,
      "grad_norm": 21.24539554413327,
      "kl": 6.5703125,
      "learning_rate": 1.899991334460036e-05,
      "loss": 0.3699,
      "reward": 1.4813492000102997,
      "reward_std": 1.0872852802276611,
      "rewards/accuracy_reward": 0.385416679084301,
      "rewards/reasoning_steps_reward": 0.866319477558136,
      "rewards/repetition_penalty_reward": -0.3628349155187607,
      "rewards/tag_count_reward": 0.5924479216337204,
      "step": 153
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 646.5052337646484,
      "epoch": 0.23097112860892388,
      "grad_norm": 7.586485137427004,
      "kl": 7.958984375,
      "learning_rate": 1.8976927486630252e-05,
      "loss": 0.3929,
      "reward": 1.2929440140724182,
      "reward_std": 0.9547273218631744,
      "rewards/accuracy_reward": 0.2760416716337204,
      "rewards/reasoning_steps_reward": 0.8802083432674408,
      "rewards/repetition_penalty_reward": -0.4609622582793236,
      "rewards/tag_count_reward": 0.5976562649011612,
      "step": 154
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 579.0521087646484,
      "epoch": 0.23247094113235844,
      "grad_norm": 5.98725832415842,
      "kl": 15.900390625,
      "learning_rate": 1.8953694699248193e-05,
      "loss": 0.3019,
      "reward": 1.338868498802185,
      "reward_std": 1.0572403371334076,
      "rewards/accuracy_reward": 0.3541666744276881,
      "rewards/reasoning_steps_reward": 0.8315972536802292,
      "rewards/repetition_penalty_reward": -0.4198121204972267,
      "rewards/tag_count_reward": 0.5729167014360428,
      "step": 155
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 464.4948043823242,
      "epoch": 0.23397075365579303,
      "grad_norm": 1.5100410826543946,
      "kl": 1.6201171875,
      "learning_rate": 1.893021562152122e-05,
      "loss": 0.2142,
      "reward": 1.4770875573158264,
      "reward_std": 1.0544297099113464,
      "rewards/accuracy_reward": 0.3541666716337204,
      "rewards/reasoning_steps_reward": 0.8368055671453476,
      "rewards/repetition_penalty_reward": -0.290707603096962,
      "rewards/tag_count_reward": 0.5768229365348816,
      "step": 156
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 352.2135543823242,
      "epoch": 0.2354705661792276,
      "grad_norm": 2.5164021567299324,
      "kl": 0.8330078125,
      "learning_rate": 1.8906490899291125e-05,
      "loss": 0.1355,
      "reward": 1.7763201594352722,
      "reward_std": 0.8412696719169617,
      "rewards/accuracy_reward": 0.4270833358168602,
      "rewards/reasoning_steps_reward": 0.869791716337204,
      "rewards/repetition_penalty_reward": -0.18852365016937256,
      "rewards/tag_count_reward": 0.6679687649011612,
      "step": 157
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 405.21875762939453,
      "epoch": 0.23697037870266216,
      "grad_norm": 1.8327817177420191,
      "kl": 0.39794921875,
      "learning_rate": 1.888252118515666e-05,
      "loss": 0.1901,
      "reward": 2.067378491163254,
      "reward_std": 0.7569487392902374,
      "rewards/accuracy_reward": 0.5468750149011612,
      "rewards/reasoning_steps_reward": 0.9409722536802292,
      "rewards/repetition_penalty_reward": -0.20302094891667366,
      "rewards/tag_count_reward": 0.7825520932674408,
      "step": 158
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 394.5520935058594,
      "epoch": 0.23847019122609675,
      "grad_norm": 1.1312490111142588,
      "kl": 0.412109375,
      "learning_rate": 1.88583071384556e-05,
      "loss": 0.0986,
      "reward": 2.053210973739624,
      "reward_std": 0.7719367817044258,
      "rewards/accuracy_reward": 0.5000000149011612,
      "rewards/reasoning_steps_reward": 0.9201389104127884,
      "rewards/repetition_penalty_reward": -0.18854257836937904,
      "rewards/tag_count_reward": 0.821614608168602,
      "step": 159
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 534.4843826293945,
      "epoch": 0.2399700037495313,
      "grad_norm": 5.729314262683275,
      "kl": 0.576171875,
      "learning_rate": 1.883384942524661e-05,
      "loss": 0.3881,
      "reward": 1.6979016363620758,
      "reward_std": 1.0777118504047394,
      "rewards/accuracy_reward": 0.3333333432674408,
      "rewards/reasoning_steps_reward": 0.8663194626569748,
      "rewards/repetition_penalty_reward": -0.2621678523719311,
      "rewards/tag_count_reward": 0.7604166716337204,
      "step": 160
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 646.6614837646484,
      "epoch": 0.24146981627296588,
      "grad_norm": 2.928392277669671,
      "kl": 0.74658203125,
      "learning_rate": 1.880914871829092e-05,
      "loss": 0.2085,
      "reward": 2.010637640953064,
      "reward_std": 0.8601251840591431,
      "rewards/accuracy_reward": 0.526041679084301,
      "rewards/reasoning_steps_reward": 0.8993055820465088,
      "rewards/repetition_penalty_reward": -0.1764284037053585,
      "rewards/tag_count_reward": 0.7617187798023224,
      "step": 161
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 804.2083587646484,
      "epoch": 0.24296962879640044,
      "grad_norm": 1.389994708780378,
      "kl": 0.5537109375,
      "learning_rate": 1.8784205697033803e-05,
      "loss": 0.0807,
      "reward": 1.8883300125598907,
      "reward_std": 0.7057935371994972,
      "rewards/accuracy_reward": 0.4479166716337204,
      "rewards/reasoning_steps_reward": 0.8576389253139496,
      "rewards/repetition_penalty_reward": -0.09170474670827389,
      "rewards/tag_count_reward": 0.6744791865348816,
      "step": 162
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 766.2187652587891,
      "epoch": 0.24446944131983503,
      "grad_norm": 5.926627814628075,
      "kl": 0.48681640625,
      "learning_rate": 1.875902104758592e-05,
      "loss": 0.0925,
      "reward": 2.008990705013275,
      "reward_std": 0.7761075049638748,
      "rewards/accuracy_reward": 0.5885416828095913,
      "rewards/reasoning_steps_reward": 0.854166716337204,
      "rewards/repetition_penalty_reward": -0.07304057851433754,
      "rewards/tag_count_reward": 0.6393229365348816,
      "step": 163
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 725.9896087646484,
      "epoch": 0.2459692538432696,
      "grad_norm": 3252.328387823972,
      "kl": 6.70703125,
      "learning_rate": 1.873359546270442e-05,
      "loss": 0.4991,
      "reward": 1.9295026063919067,
      "reward_std": 0.8431710749864578,
      "rewards/accuracy_reward": 0.5833333507180214,
      "rewards/reasoning_steps_reward": 0.8489583283662796,
      "rewards/repetition_penalty_reward": -0.0731016006320715,
      "rewards/tag_count_reward": 0.5703125298023224,
      "step": 164
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 711.4948120117188,
      "epoch": 0.24746906636670415,
      "grad_norm": 35.24975659123428,
      "kl": 0.7900390625,
      "learning_rate": 1.8707929641773876e-05,
      "loss": 0.2004,
      "reward": 1.7060418128967285,
      "reward_std": 0.8163315951824188,
      "rewards/accuracy_reward": 0.36458334140479565,
      "rewards/reasoning_steps_reward": 0.8437500447034836,
      "rewards/repetition_penalty_reward": -0.06739573087543249,
      "rewards/tag_count_reward": 0.5651041865348816,
      "step": 165
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 631.3385772705078,
      "epoch": 0.24896887889013875,
      "grad_norm": 44.81319877908832,
      "kl": 0.48486328125,
      "learning_rate": 1.8682024290787092e-05,
      "loss": 0.2422,
      "reward": 1.8670725524425507,
      "reward_std": 0.7701267302036285,
      "rewards/accuracy_reward": 0.5000000149011612,
      "rewards/reasoning_steps_reward": 0.8663195073604584,
      "rewards/repetition_penalty_reward": -0.07086153514683247,
      "rewards/tag_count_reward": 0.571614608168602,
      "step": 166
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 610.9791870117188,
      "epoch": 0.2504686914135733,
      "grad_norm": 48.22852108204421,
      "kl": 1.1845703125,
      "learning_rate": 1.8655880122325633e-05,
      "loss": 0.2746,
      "reward": 1.8831318020820618,
      "reward_std": 0.737670287489891,
      "rewards/accuracy_reward": 0.4895833432674408,
      "rewards/reasoning_steps_reward": 0.8975694924592972,
      "rewards/repetition_penalty_reward": -0.07042735442519188,
      "rewards/tag_count_reward": 0.5664062798023224,
      "step": 167
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 637.8437805175781,
      "epoch": 0.25196850393700787,
      "grad_norm": 127.32727674691347,
      "kl": 2.96484375,
      "learning_rate": 1.862949785554025e-05,
      "loss": 0.4427,
      "reward": 1.6282338500022888,
      "reward_std": 0.8880203366279602,
      "rewards/accuracy_reward": 0.4114583432674408,
      "rewards/reasoning_steps_reward": 0.7916667014360428,
      "rewards/repetition_penalty_reward": -0.05535993352532387,
      "rewards/tag_count_reward": 0.4804687649011612,
      "step": 168
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 593.1562652587891,
      "epoch": 0.25346831646044243,
      "grad_norm": 66.6798484485283,
      "kl": 2.54296875,
      "learning_rate": 1.8602878216131093e-05,
      "loss": 0.4588,
      "reward": 1.8000280857086182,
      "reward_std": 0.8398674130439758,
      "rewards/accuracy_reward": 0.4739583432674408,
      "rewards/reasoning_steps_reward": 0.8541666865348816,
      "rewards/repetition_penalty_reward": -0.059346938505768776,
      "rewards/tag_count_reward": 0.5312500149011612,
      "step": 169
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 600.9687652587891,
      "epoch": 0.254968128983877,
      "grad_norm": 1723.0756525306333,
      "kl": 5.98046875,
      "learning_rate": 1.8576021936327747e-05,
      "loss": 0.6964,
      "reward": 1.6876211762428284,
      "reward_std": 0.9227285087108612,
      "rewards/accuracy_reward": 0.479166679084301,
      "rewards/reasoning_steps_reward": 0.7777778059244156,
      "rewards/repetition_penalty_reward": -0.04718789644539356,
      "rewards/tag_count_reward": 0.4778645932674408,
      "step": 170
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 570.3333435058594,
      "epoch": 0.25646794150731156,
      "grad_norm": 1469.676229468339,
      "kl": 11.125,
      "learning_rate": 1.8548929754869095e-05,
      "loss": 1.1501,
      "reward": 1.6916078925132751,
      "reward_std": 0.8378966599702835,
      "rewards/accuracy_reward": 0.4687500149011612,
      "rewards/reasoning_steps_reward": 0.7812500298023224,
      "rewards/repetition_penalty_reward": -0.050579698756337166,
      "rewards/tag_count_reward": 0.4921875149011612,
      "step": 171
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 634.5416870117188,
      "epoch": 0.2579677540307462,
      "grad_norm": 308.20087481858275,
      "kl": 15.05859375,
      "learning_rate": 1.8521602416982998e-05,
      "loss": 0.8609,
      "reward": 1.5223284363746643,
      "reward_std": 0.8932169824838638,
      "rewards/accuracy_reward": 0.3906250149011612,
      "rewards/reasoning_steps_reward": 0.7343750149011612,
      "rewards/repetition_penalty_reward": -0.044077920727431774,
      "rewards/tag_count_reward": 0.4414062649011612,
      "step": 172
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 575.4375228881836,
      "epoch": 0.25946756655418074,
      "grad_norm": 27.514037083049207,
      "kl": 1.0087890625,
      "learning_rate": 1.8494040674365785e-05,
      "loss": 0.4806,
      "reward": 1.7090361416339874,
      "reward_std": 0.9580790549516678,
      "rewards/accuracy_reward": 0.5000000223517418,
      "rewards/reasoning_steps_reward": 0.7500000298023224,
      "rewards/repetition_penalty_reward": -0.03966184053570032,
      "rewards/tag_count_reward": 0.4986979365348816,
      "step": 173
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 540.9948043823242,
      "epoch": 0.2609673790776153,
      "grad_norm": 24.242426884849852,
      "kl": 0.8251953125,
      "learning_rate": 1.8466245285161593e-05,
      "loss": 0.4027,
      "reward": 1.7482682466506958,
      "reward_std": 0.8416074514389038,
      "rewards/accuracy_reward": 0.4270833432674408,
      "rewards/reasoning_steps_reward": 0.8246528506278992,
      "rewards/repetition_penalty_reward": -0.04773869924247265,
      "rewards/tag_count_reward": 0.544270858168602,
      "step": 174
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 527.1041717529297,
      "epoch": 0.26246719160104987,
      "grad_norm": 13.61888783625401,
      "kl": 1.12109375,
      "learning_rate": 1.8438217013941494e-05,
      "loss": 0.4855,
      "reward": 1.8302274644374847,
      "reward_std": 0.9374262690544128,
      "rewards/accuracy_reward": 0.4375000149011612,
      "rewards/reasoning_steps_reward": 0.8263889253139496,
      "rewards/repetition_penalty_reward": -0.05215102527290583,
      "rewards/tag_count_reward": 0.618489608168602,
      "step": 175
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 509.71356201171875,
      "epoch": 0.26396700412448443,
      "grad_norm": 9.502433027265221,
      "kl": 0.9228515625,
      "learning_rate": 1.8409956631682475e-05,
      "loss": 0.5068,
      "reward": 1.8384932577610016,
      "reward_std": 0.9605595469474792,
      "rewards/accuracy_reward": 0.4270833507180214,
      "rewards/reasoning_steps_reward": 0.80555559694767,
      "rewards/repetition_penalty_reward": -0.03607274405658245,
      "rewards/tag_count_reward": 0.641927108168602,
      "step": 176
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 542.1718902587891,
      "epoch": 0.265466816647919,
      "grad_norm": 12.713070926311392,
      "kl": 1.177734375,
      "learning_rate": 1.838146491574624e-05,
      "loss": 0.5857,
      "reward": 1.5894387364387512,
      "reward_std": 1.0209160447120667,
      "rewards/accuracy_reward": 0.4114583507180214,
      "rewards/reasoning_steps_reward": 0.6440972313284874,
      "rewards/repetition_penalty_reward": -0.02991898776963353,
      "rewards/tag_count_reward": 0.5638020932674408,
      "step": 177
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 547.8333435058594,
      "epoch": 0.26696662917135355,
      "grad_norm": 16.206516622301223,
      "kl": 1.232421875,
      "learning_rate": 1.83527426498578e-05,
      "loss": 0.6041,
      "reward": 1.6392149925231934,
      "reward_std": 1.0323166698217392,
      "rewards/accuracy_reward": 0.3281250037252903,
      "rewards/reasoning_steps_reward": 0.756944477558136,
      "rewards/repetition_penalty_reward": -0.03830242808908224,
      "rewards/tag_count_reward": 0.5924479365348816,
      "step": 178
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 536.1562805175781,
      "epoch": 0.26846644169478817,
      "grad_norm": 129.17932118885184,
      "kl": 3.412109375,
      "learning_rate": 1.832379062408394e-05,
      "loss": 0.7632,
      "reward": 1.5638504028320312,
      "reward_std": 0.9709765613079071,
      "rewards/accuracy_reward": 0.3229166716337204,
      "rewards/reasoning_steps_reward": 0.6875000298023224,
      "rewards/repetition_penalty_reward": -0.03120174235664308,
      "rewards/tag_count_reward": 0.5846354216337204,
      "step": 179
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 536.7291870117188,
      "epoch": 0.26996625421822273,
      "grad_norm": 60.94992386504032,
      "kl": 1.783203125,
      "learning_rate": 1.8294609634811482e-05,
      "loss": 0.6702,
      "reward": 1.7184252440929413,
      "reward_std": 1.1222001910209656,
      "rewards/accuracy_reward": 0.4687500223517418,
      "rewards/reasoning_steps_reward": 0.671875,
      "rewards/repetition_penalty_reward": -0.02636647690087557,
      "rewards/tag_count_reward": 0.6041666716337204,
      "step": 180
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 488.0625228881836,
      "epoch": 0.2714660667416573,
      "grad_norm": 597.8726346831836,
      "kl": 4.18359375,
      "learning_rate": 1.8265200484725364e-05,
      "loss": 0.882,
      "reward": 1.7504200041294098,
      "reward_std": 1.0297911912202835,
      "rewards/accuracy_reward": 0.4270833507180214,
      "rewards/reasoning_steps_reward": 0.7291666865348816,
      "rewards/repetition_penalty_reward": -0.03213207516819239,
      "rewards/tag_count_reward": 0.626302108168602,
      "step": 181
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 577.8541870117188,
      "epoch": 0.27296587926509186,
      "grad_norm": 289.5259006408696,
      "kl": 3.12109375,
      "learning_rate": 1.823556398278657e-05,
      "loss": 0.7697,
      "reward": 1.4063380062580109,
      "reward_std": 1.146880030632019,
      "rewards/accuracy_reward": 0.3437500037252903,
      "rewards/reasoning_steps_reward": 0.5381944701075554,
      "rewards/repetition_penalty_reward": -0.027689829003065825,
      "rewards/tag_count_reward": 0.5520833507180214,
      "step": 182
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 544.1302337646484,
      "epoch": 0.2744656917885264,
      "grad_norm": 18278.821342314448,
      "kl": 62.75,
      "learning_rate": 1.820570094420989e-05,
      "loss": 5.0174,
      "reward": 1.5145381093025208,
      "reward_std": 1.0834265649318695,
      "rewards/accuracy_reward": 0.3333333432674408,
      "rewards/reasoning_steps_reward": 0.6649305820465088,
      "rewards/repetition_penalty_reward": -0.03320494340732694,
      "rewards/tag_count_reward": 0.5494791865348816,
      "step": 183
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 528.7239761352539,
      "epoch": 0.275965504311961,
      "grad_norm": 115.68017983054848,
      "kl": 2.783203125,
      "learning_rate": 1.817561219044148e-05,
      "loss": 0.7618,
      "reward": 1.561537265777588,
      "reward_std": 1.062522441148758,
      "rewards/accuracy_reward": 0.3489583469927311,
      "rewards/reasoning_steps_reward": 0.6597222089767456,
      "rewards/repetition_penalty_reward": -0.033080867026001215,
      "rewards/tag_count_reward": 0.5859375149011612,
      "step": 184
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 552.9062652587891,
      "epoch": 0.27746531683539555,
      "grad_norm": 185.26777317096048,
      "kl": 14.705078125,
      "learning_rate": 1.814529854913626e-05,
      "loss": 0.8482,
      "reward": 1.5542615354061127,
      "reward_std": 1.0569299161434174,
      "rewards/accuracy_reward": 0.3281250074505806,
      "rewards/reasoning_steps_reward": 0.6510417014360428,
      "rewards/repetition_penalty_reward": -0.03297804016619921,
      "rewards/tag_count_reward": 0.6080729365348816,
      "step": 185
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 569.4687805175781,
      "epoch": 0.27896512935883017,
      "grad_norm": 711.4722614182526,
      "kl": 4.55859375,
      "learning_rate": 1.8114760854135168e-05,
      "loss": 0.8749,
      "reward": 1.4769874811172485,
      "reward_std": 1.1705779433250427,
      "rewards/accuracy_reward": 0.3489583469927311,
      "rewards/reasoning_steps_reward": 0.605902835726738,
      "rewards/repetition_penalty_reward": -0.024748651776462793,
      "rewards/tag_count_reward": 0.5468750298023224,
      "step": 186
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 580.2656402587891,
      "epoch": 0.28046494188226473,
      "grad_norm": 3683.741373123422,
      "kl": 6.32421875,
      "learning_rate": 1.808399994544222e-05,
      "loss": 0.9344,
      "reward": 1.4767035245895386,
      "reward_std": 1.1781282722949982,
      "rewards/accuracy_reward": 0.3645833432674408,
      "rewards/reasoning_steps_reward": 0.5954861342906952,
      "rewards/repetition_penalty_reward": -0.02503264555707574,
      "rewards/tag_count_reward": 0.5416666939854622,
      "step": 187
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 512.8333511352539,
      "epoch": 0.2819647544056993,
      "grad_norm": 231.95111964687146,
      "kl": 160.20703125,
      "learning_rate": 1.805301666920138e-05,
      "loss": 0.9703,
      "reward": 1.5358233749866486,
      "reward_std": 1.097620114684105,
      "rewards/accuracy_reward": 0.3125000074505806,
      "rewards/reasoning_steps_reward": 0.6510417014360428,
      "rewards/repetition_penalty_reward": -0.03318707179278135,
      "rewards/tag_count_reward": 0.6054687798023224,
      "step": 188
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 532.8073120117188,
      "epoch": 0.28346456692913385,
      "grad_norm": 392.6089381205205,
      "kl": 4.78515625,
      "learning_rate": 1.802181187767332e-05,
      "loss": 0.8115,
      "reward": 1.4347640872001648,
      "reward_std": 1.0974180102348328,
      "rewards/accuracy_reward": 0.291666679084301,
      "rewards/reasoning_steps_reward": 0.6093750298023224,
      "rewards/repetition_penalty_reward": -0.028777593281120062,
      "rewards/tag_count_reward": 0.5625000149011612,
      "step": 189
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 590.7552185058594,
      "epoch": 0.2849643794525684,
      "grad_norm": 230.8284538739948,
      "kl": 13.0234375,
      "learning_rate": 1.7990386429211945e-05,
      "loss": 1.0111,
      "reward": 1.3888815939426422,
      "reward_std": 1.1380842924118042,
      "rewards/accuracy_reward": 0.2968750074505806,
      "rewards/reasoning_steps_reward": 0.5781250298023224,
      "rewards/repetition_penalty_reward": -0.02518094959668815,
      "rewards/tag_count_reward": 0.5390625149011612,
      "step": 190
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 523.4375076293945,
      "epoch": 0.286464191976003,
      "grad_norm": 198.59410394134304,
      "kl": 2.521484375,
      "learning_rate": 1.7958741188240808e-05,
      "loss": 0.763,
      "reward": 1.4029823541641235,
      "reward_std": 1.0676278173923492,
      "rewards/accuracy_reward": 0.2864583469927311,
      "rewards/reasoning_steps_reward": 0.5659722536802292,
      "rewards/repetition_penalty_reward": -0.030177415814250708,
      "rewards/tag_count_reward": 0.5807291865348816,
      "step": 191
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 533.6666793823242,
      "epoch": 0.2879640044994376,
      "grad_norm": 711.4562338474254,
      "kl": 26.0546875,
      "learning_rate": 1.7926877025229313e-05,
      "loss": 1.868,
      "reward": 1.5179384350776672,
      "reward_std": 1.1407716572284698,
      "rewards/accuracy_reward": 0.3437500074505806,
      "rewards/reasoning_steps_reward": 0.6128472536802292,
      "rewards/repetition_penalty_reward": -0.02720050560310483,
      "rewards/tag_count_reward": 0.5885416865348816,
      "step": 192
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 532.0208511352539,
      "epoch": 0.28946381702287216,
      "grad_norm": 68.3445984974822,
      "kl": 1.984375,
      "learning_rate": 1.789479481666878e-05,
      "loss": 0.7144,
      "reward": 1.4272576868534088,
      "reward_std": 1.1221881210803986,
      "rewards/accuracy_reward": 0.2916666753590107,
      "rewards/reasoning_steps_reward": 0.5798611342906952,
      "rewards/repetition_penalty_reward": -0.021093112416565418,
      "rewards/tag_count_reward": 0.5768229216337204,
      "step": 193
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 591.0364761352539,
      "epoch": 0.2909636295463067,
      "grad_norm": 50.22604651230194,
      "kl": 2.06640625,
      "learning_rate": 1.786249544504834e-05,
      "loss": 0.6381,
      "reward": 1.2664842307567596,
      "reward_std": 1.0500086545944214,
      "rewards/accuracy_reward": 0.24479167722165585,
      "rewards/reasoning_steps_reward": 0.5451388955116272,
      "rewards/repetition_penalty_reward": -0.024748508352786303,
      "rewards/tag_count_reward": 0.5013020932674408,
      "step": 194
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 491.8958435058594,
      "epoch": 0.2924634420697413,
      "grad_norm": 133.86316178041838,
      "kl": 3.8955078125,
      "learning_rate": 1.7829979798830646e-05,
      "loss": 0.7041,
      "reward": 1.6353325545787811,
      "reward_std": 1.0081156641244888,
      "rewards/accuracy_reward": 0.3750000074505806,
      "rewards/reasoning_steps_reward": 0.6684028208255768,
      "rewards/repetition_penalty_reward": -0.039580670185387135,
      "rewards/tag_count_reward": 0.631510429084301,
      "step": 195
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 519.8072967529297,
      "epoch": 0.29396325459317585,
      "grad_norm": 44.76317495895618,
      "kl": 3.6826171875,
      "learning_rate": 1.779724877242745e-05,
      "loss": 0.6891,
      "reward": 1.4282979369163513,
      "reward_std": 1.150389850139618,
      "rewards/accuracy_reward": 0.2968750111758709,
      "rewards/reasoning_steps_reward": 0.5972222685813904,
      "rewards/repetition_penalty_reward": -0.051736912690103054,
      "rewards/tag_count_reward": 0.5859375149011612,
      "step": 196
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 498.42708587646484,
      "epoch": 0.2954630671166104,
      "grad_norm": 48.296982691961496,
      "kl": 0.994140625,
      "learning_rate": 1.776430326617498e-05,
      "loss": 0.6805,
      "reward": 1.6272485554218292,
      "reward_std": 1.1874802112579346,
      "rewards/accuracy_reward": 0.4166666716337204,
      "rewards/reasoning_steps_reward": 0.6458333432674408,
      "rewards/repetition_penalty_reward": -0.07327230088412762,
      "rewards/tag_count_reward": 0.6380208432674408,
      "step": 197
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 516.9583435058594,
      "epoch": 0.296962879640045,
      "grad_norm": 70.18509930572817,
      "kl": 2.08203125,
      "learning_rate": 1.77311441863092e-05,
      "loss": 0.7866,
      "reward": 1.406771183013916,
      "reward_std": 1.206770658493042,
      "rewards/accuracy_reward": 0.3229166716337204,
      "rewards/reasoning_steps_reward": 0.5833333730697632,
      "rewards/repetition_penalty_reward": -0.0867184977978468,
      "rewards/tag_count_reward": 0.587239608168602,
      "step": 198
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 543.0312652587891,
      "epoch": 0.2984626921634796,
      "grad_norm": 45.26003893019006,
      "kl": 1.7421875,
      "learning_rate": 1.769777244494086e-05,
      "loss": 0.6958,
      "reward": 1.30002062022686,
      "reward_std": 1.114796221256256,
      "rewards/accuracy_reward": 0.2447916753590107,
      "rewards/reasoning_steps_reward": 0.569444477558136,
      "rewards/repetition_penalty_reward": -0.08583013713359833,
      "rewards/tag_count_reward": 0.5716145932674408,
      "step": 199
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 517.4166793823242,
      "epoch": 0.29996250468691416,
      "grad_norm": 29.376902334980826,
      "kl": 1.3662109375,
      "learning_rate": 1.7664188960030422e-05,
      "loss": 0.6637,
      "reward": 1.3883242011070251,
      "reward_std": 1.100932002067566,
      "rewards/accuracy_reward": 0.2552083432674408,
      "rewards/reasoning_steps_reward": 0.5937500149011612,
      "rewards/repetition_penalty_reward": -0.04136331286281347,
      "rewards/tag_count_reward": 0.5807291865348816,
      "step": 200
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 550.4062652587891,
      "epoch": 0.3014623172103487,
      "grad_norm": 17.485691153105112,
      "kl": 1.34765625,
      "learning_rate": 1.7630394655362798e-05,
      "loss": 0.6612,
      "reward": 1.3975980877876282,
      "reward_std": 1.0539609044790268,
      "rewards/accuracy_reward": 0.322916679084301,
      "rewards/reasoning_steps_reward": 0.5434028059244156,
      "rewards/repetition_penalty_reward": -0.031221389304846525,
      "rewards/tag_count_reward": 0.5625000149011612,
      "step": 201
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 527.9635543823242,
      "epoch": 0.3029621297337833,
      "grad_norm": 90.79232868194525,
      "kl": 1.744140625,
      "learning_rate": 1.7596390460521946e-05,
      "loss": 0.7488,
      "reward": 1.4379164278507233,
      "reward_std": 1.1352272033691406,
      "rewards/accuracy_reward": 0.3020833432674408,
      "rewards/reasoning_steps_reward": 0.5781250149011612,
      "rewards/repetition_penalty_reward": -0.025625293143093586,
      "rewards/tag_count_reward": 0.5833333432674408,
      "step": 202
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 531.1771087646484,
      "epoch": 0.30446194225721784,
      "grad_norm": 7199.74671681739,
      "kl": 48.05078125,
      "learning_rate": 1.7562177310865296e-05,
      "loss": 4.0565,
      "reward": 1.4444026350975037,
      "reward_std": 1.1348033249378204,
      "rewards/accuracy_reward": 0.3072916716337204,
      "rewards/reasoning_steps_reward": 0.57118059694767,
      "rewards/repetition_penalty_reward": -0.033027936704456806,
      "rewards/tag_count_reward": 0.598958358168602,
      "step": 203
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 506.8073043823242,
      "epoch": 0.3059617547806524,
      "grad_norm": 1098.2572442297885,
      "kl": 54.375,
      "learning_rate": 1.7527756147498026e-05,
      "loss": 4.475,
      "reward": 1.5475478768348694,
      "reward_std": 1.118485450744629,
      "rewards/accuracy_reward": 0.3645833432674408,
      "rewards/reasoning_steps_reward": 0.6041667014360428,
      "rewards/repetition_penalty_reward": -0.03187928069382906,
      "rewards/tag_count_reward": 0.6106770932674408,
      "step": 204
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 506.47396087646484,
      "epoch": 0.30746156730408697,
      "grad_norm": 257.4360594846129,
      "kl": 16.2890625,
      "learning_rate": 1.7493127917247168e-05,
      "loss": 1.7909,
      "reward": 1.476582944393158,
      "reward_std": 1.1201183497905731,
      "rewards/accuracy_reward": 0.3125000074505806,
      "rewards/reasoning_steps_reward": 0.60243059694767,
      "rewards/repetition_penalty_reward": -0.0255872611887753,
      "rewards/tag_count_reward": 0.587239608168602,
      "step": 205
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 474.36981201171875,
      "epoch": 0.3089613798275216,
      "grad_norm": 82.09981426436468,
      "kl": 20.0625,
      "learning_rate": 1.7458293572635573e-05,
      "loss": 0.7928,
      "reward": 1.5302923321723938,
      "reward_std": 1.1175011098384857,
      "rewards/accuracy_reward": 0.3177083432674408,
      "rewards/reasoning_steps_reward": 0.625,
      "rewards/repetition_penalty_reward": -0.03741603484377265,
      "rewards/tag_count_reward": 0.6250000149011612,
      "step": 206
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 488.75000762939453,
      "epoch": 0.31046119235095615,
      "grad_norm": 96.44909255545524,
      "kl": 2.359375,
      "learning_rate": 1.7423254071855696e-05,
      "loss": 0.7126,
      "reward": 1.6391958892345428,
      "reward_std": 1.2212097346782684,
      "rewards/accuracy_reward": 0.4375000074505806,
      "rewards/reasoning_steps_reward": 0.6250000298023224,
      "rewards/repetition_penalty_reward": -0.0313770417124033,
      "rewards/tag_count_reward": 0.6080729365348816,
      "step": 207
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 434.65626525878906,
      "epoch": 0.3119610048743907,
      "grad_norm": 120.67164118704962,
      "kl": 24.044921875,
      "learning_rate": 1.7388010378743255e-05,
      "loss": 0.8024,
      "reward": 1.6531369984149933,
      "reward_std": 1.085878610610962,
      "rewards/accuracy_reward": 0.3750000074505806,
      "rewards/reasoning_steps_reward": 0.6770833432674408,
      "rewards/repetition_penalty_reward": -0.03696717880666256,
      "rewards/tag_count_reward": 0.638020858168602,
      "step": 208
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 416.17189025878906,
      "epoch": 0.3134608173978253,
      "grad_norm": 106.78168334463734,
      "kl": 3.224609375,
      "learning_rate": 1.735256346275071e-05,
      "loss": 0.8412,
      "reward": 1.7143349051475525,
      "reward_std": 1.07014599442482,
      "rewards/accuracy_reward": 0.3958333507180214,
      "rewards/reasoning_steps_reward": 0.677083358168602,
      "rewards/repetition_penalty_reward": -0.029154742136597633,
      "rewards/tag_count_reward": 0.6705729365348816,
      "step": 209
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 448.38021087646484,
      "epoch": 0.31496062992125984,
      "grad_norm": 133.47773562969348,
      "kl": 3.859375,
      "learning_rate": 1.7316914298920592e-05,
      "loss": 0.983,
      "reward": 1.591974526643753,
      "reward_std": 1.1013777256011963,
      "rewards/accuracy_reward": 0.354166679084301,
      "rewards/reasoning_steps_reward": 0.63368059694767,
      "rewards/repetition_penalty_reward": -0.032591511495411396,
      "rewards/tag_count_reward": 0.6367187649011612,
      "step": 210
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 436.5364685058594,
      "epoch": 0.3164604424446944,
      "grad_norm": 108.63670939149665,
      "kl": 4.8046875,
      "learning_rate": 1.7281063867858687e-05,
      "loss": 0.9645,
      "reward": 1.6065455377101898,
      "reward_std": 1.0716453790664673,
      "rewards/accuracy_reward": 0.3593750149011612,
      "rewards/reasoning_steps_reward": 0.657986119389534,
      "rewards/repetition_penalty_reward": -0.04493015632033348,
      "rewards/tag_count_reward": 0.6341145932674408,
      "step": 211
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 398.9948043823242,
      "epoch": 0.31796025496812896,
      "grad_norm": 96.14698671429733,
      "kl": 83.048828125,
      "learning_rate": 1.7245013155707076e-05,
      "loss": 0.8663,
      "reward": 1.7979373633861542,
      "reward_std": 1.0779232680797577,
      "rewards/accuracy_reward": 0.4375,
      "rewards/reasoning_steps_reward": 0.7152777910232544,
      "rewards/repetition_penalty_reward": -0.034527900628745556,
      "rewards/tag_count_reward": 0.6796875149011612,
      "step": 212
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 443.6145935058594,
      "epoch": 0.3194600674915636,
      "grad_norm": 409.81304589430863,
      "kl": 6.1015625,
      "learning_rate": 1.7208763154116973e-05,
      "loss": 1.0656,
      "reward": 1.603056788444519,
      "reward_std": 1.083729773759842,
      "rewards/accuracy_reward": 0.3750000074505806,
      "rewards/reasoning_steps_reward": 0.645833358168602,
      "rewards/repetition_penalty_reward": -0.031057825777679682,
      "rewards/tag_count_reward": 0.6132812798023224,
      "step": 213
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 461.0364685058594,
      "epoch": 0.32095988001499814,
      "grad_norm": 57.764921292245205,
      "kl": 1.998046875,
      "learning_rate": 1.7172314860221494e-05,
      "loss": 0.7731,
      "reward": 1.5737330913543701,
      "reward_std": 1.0800977945327759,
      "rewards/accuracy_reward": 0.3229166865348816,
      "rewards/reasoning_steps_reward": 0.64930559694767,
      "rewards/repetition_penalty_reward": -0.03781210444867611,
      "rewards/tag_count_reward": 0.6393229365348816,
      "step": 214
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 478.05731201171875,
      "epoch": 0.3224596925384327,
      "grad_norm": 50.79407157857832,
      "kl": 7.47265625,
      "learning_rate": 1.713566927660818e-05,
      "loss": 0.7083,
      "reward": 1.4539363086223602,
      "reward_std": 1.0611682534217834,
      "rewards/accuracy_reward": 0.2447916679084301,
      "rewards/reasoning_steps_reward": 0.6440972536802292,
      "rewards/repetition_penalty_reward": -0.03391093295067549,
      "rewards/tag_count_reward": 0.5989583432674408,
      "step": 215
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 418.6614761352539,
      "epoch": 0.32395950506186727,
      "grad_norm": 62.35489423972326,
      "kl": 1.40625,
      "learning_rate": 1.7098827411291474e-05,
      "loss": 0.7932,
      "reward": 1.7967810034751892,
      "reward_std": 1.071500152349472,
      "rewards/accuracy_reward": 0.5052083432674408,
      "rewards/reasoning_steps_reward": 0.6666667014360428,
      "rewards/repetition_penalty_reward": -0.0443648905493319,
      "rewards/tag_count_reward": 0.6692708432674408,
      "step": 216
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 465.15625762939453,
      "epoch": 0.32545931758530183,
      "grad_norm": 523.1686817311862,
      "kl": 14.51171875,
      "learning_rate": 1.7061790277684935e-05,
      "loss": 1.1443,
      "reward": 1.7601740062236786,
      "reward_std": 0.9184492155909538,
      "rewards/accuracy_reward": 0.463541679084301,
      "rewards/reasoning_steps_reward": 0.6875000298023224,
      "rewards/repetition_penalty_reward": -0.03670105990022421,
      "rewards/tag_count_reward": 0.645833358168602,
      "step": 217
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 440.9166793823242,
      "epoch": 0.3269591301087364,
      "grad_norm": 138.26908034820204,
      "kl": 2.8447265625,
      "learning_rate": 1.7024558894573408e-05,
      "loss": 0.8425,
      "reward": 1.6384324729442596,
      "reward_std": 1.0819440335035324,
      "rewards/accuracy_reward": 0.3750000074505806,
      "rewards/reasoning_steps_reward": 0.675347238779068,
      "rewards/repetition_penalty_reward": -0.035612753592431545,
      "rewards/tag_count_reward": 0.6236979365348816,
      "step": 218
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 430.1927185058594,
      "epoch": 0.32845894263217096,
      "grad_norm": 36.752519571585786,
      "kl": 1.0751953125,
      "learning_rate": 1.698713428608497e-05,
      "loss": 0.6927,
      "reward": 1.7269428670406342,
      "reward_std": 1.1123250424861908,
      "rewards/accuracy_reward": 0.4427083507180214,
      "rewards/reasoning_steps_reward": 0.6840278208255768,
      "rewards/repetition_penalty_reward": -0.036512063816189766,
      "rewards/tag_count_reward": 0.6367187798023224,
      "step": 219
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 428.6510467529297,
      "epoch": 0.3299587551556056,
      "grad_norm": 27.095258985089583,
      "kl": 1.2314453125,
      "learning_rate": 1.694951748166278e-05,
      "loss": 0.6542,
      "reward": 1.7653653621673584,
      "reward_std": 1.0500112771987915,
      "rewards/accuracy_reward": 0.4270833432674408,
      "rewards/reasoning_steps_reward": 0.734375,
      "rewards/repetition_penalty_reward": -0.0458325962536037,
      "rewards/tag_count_reward": 0.649739608168602,
      "step": 220
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 413.37500762939453,
      "epoch": 0.33145856767904014,
      "grad_norm": 47.16742116134022,
      "kl": 2.20703125,
      "learning_rate": 1.6911709516036755e-05,
      "loss": 0.7411,
      "reward": 1.8477334082126617,
      "reward_std": 1.0211681723594666,
      "rewards/accuracy_reward": 0.4427083432674408,
      "rewards/reasoning_steps_reward": 0.769097238779068,
      "rewards/repetition_penalty_reward": -0.04245765320956707,
      "rewards/tag_count_reward": 0.6783854365348816,
      "step": 221
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 397.3645935058594,
      "epoch": 0.3329583802024747,
      "grad_norm": 45.5053118676597,
      "kl": 2.3515625,
      "learning_rate": 1.6873711429195095e-05,
      "loss": 0.8572,
      "reward": 1.7332542836666107,
      "reward_std": 1.0097034871578217,
      "rewards/accuracy_reward": 0.3906250074505806,
      "rewards/reasoning_steps_reward": 0.7309028059244156,
      "rewards/repetition_penalty_reward": -0.03801308758556843,
      "rewards/tag_count_reward": 0.649739608168602,
      "step": 222
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 435.78126525878906,
      "epoch": 0.33445819272590926,
      "grad_norm": 2129.317338646113,
      "kl": 12.60546875,
      "learning_rate": 1.6835524266355698e-05,
      "loss": 1.8881,
      "reward": 1.8117730021476746,
      "reward_std": 0.9651193022727966,
      "rewards/accuracy_reward": 0.4687500149011612,
      "rewards/reasoning_steps_reward": 0.7083333283662796,
      "rewards/repetition_penalty_reward": -0.03978960122913122,
      "rewards/tag_count_reward": 0.6744791716337204,
      "step": 223
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 315.84375762939453,
      "epoch": 0.3359580052493438,
      "grad_norm": 341.57819859952355,
      "kl": 2.9423828125,
      "learning_rate": 1.6797149077937395e-05,
      "loss": 0.5076,
      "reward": 1.869988203048706,
      "reward_std": 0.9940178692340851,
      "rewards/accuracy_reward": 0.4687500074505806,
      "rewards/reasoning_steps_reward": 0.765625,
      "rewards/repetition_penalty_reward": -0.05188690684735775,
      "rewards/tag_count_reward": 0.6875000149011612,
      "step": 224
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 330.26563262939453,
      "epoch": 0.3374578177727784,
      "grad_norm": 430.46296061670853,
      "kl": 2.5078125,
      "learning_rate": 1.6758586919531054e-05,
      "loss": 0.5861,
      "reward": 1.754932165145874,
      "reward_std": 0.967212975025177,
      "rewards/accuracy_reward": 0.4114583432674408,
      "rewards/reasoning_steps_reward": 0.7517361342906952,
      "rewards/repetition_penalty_reward": -0.03847070410847664,
      "rewards/tag_count_reward": 0.6302083432674408,
      "step": 225
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 345.4322929382324,
      "epoch": 0.33895763029621295,
      "grad_norm": 357.14681488285794,
      "kl": 9.0390625,
      "learning_rate": 1.671983885187055e-05,
      "loss": 0.4549,
      "reward": 1.7626317143440247,
      "reward_std": 0.9779582172632217,
      "rewards/accuracy_reward": 0.432291679084301,
      "rewards/reasoning_steps_reward": 0.7309027910232544,
      "rewards/repetition_penalty_reward": -0.035979412496089935,
      "rewards/tag_count_reward": 0.6354167014360428,
      "step": 226
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 446.6510467529297,
      "epoch": 0.34045744281964757,
      "grad_norm": 155.57552226255254,
      "kl": 1.998046875,
      "learning_rate": 1.6680905940803596e-05,
      "loss": 0.7187,
      "reward": 1.6927993595600128,
      "reward_std": 1.1329753398895264,
      "rewards/accuracy_reward": 0.4843750074505806,
      "rewards/reasoning_steps_reward": 0.6996527910232544,
      "rewards/repetition_penalty_reward": -0.04721810668706894,
      "rewards/tag_count_reward": 0.555989608168602,
      "step": 227
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 400.7708435058594,
      "epoch": 0.34195725534308213,
      "grad_norm": 52.313264872011814,
      "kl": 1.01904296875,
      "learning_rate": 1.66417892572624e-05,
      "loss": 0.6703,
      "reward": 1.7310058176517487,
      "reward_std": 0.9815861284732819,
      "rewards/accuracy_reward": 0.411458358168602,
      "rewards/reasoning_steps_reward": 0.7361111491918564,
      "rewards/repetition_penalty_reward": -0.05458456836640835,
      "rewards/tag_count_reward": 0.6380208432674408,
      "step": 228
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 362.2291793823242,
      "epoch": 0.3434570678665167,
      "grad_norm": 3.414097906304966,
      "kl": 0.44677734375,
      "learning_rate": 1.6602489877234235e-05,
      "loss": 0.5655,
      "reward": 1.9467909336090088,
      "reward_std": 0.9363191574811935,
      "rewards/accuracy_reward": 0.494791679084301,
      "rewards/reasoning_steps_reward": 0.8229166865348816,
      "rewards/repetition_penalty_reward": -0.04409462306648493,
      "rewards/tag_count_reward": 0.6731770932674408,
      "step": 229
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 382.7916793823242,
      "epoch": 0.34495688038995126,
      "grad_norm": 1.2104539856778693,
      "kl": 0.3408203125,
      "learning_rate": 1.656300888173181e-05,
      "loss": 0.4857,
      "reward": 1.8293315470218658,
      "reward_std": 0.9052053391933441,
      "rewards/accuracy_reward": 0.4375000074505806,
      "rewards/reasoning_steps_reward": 0.7795139253139496,
      "rewards/repetition_penalty_reward": -0.04393233545124531,
      "rewards/tag_count_reward": 0.6562500298023224,
      "step": 230
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 432.2864761352539,
      "epoch": 0.3464566929133858,
      "grad_norm": 2.412360731904819,
      "kl": 0.49462890625,
      "learning_rate": 1.6523347356763572e-05,
      "loss": 0.5985,
      "reward": 1.8154855072498322,
      "reward_std": 1.1332524120807648,
      "rewards/accuracy_reward": 0.5000000223517418,
      "rewards/reasoning_steps_reward": 0.723958358168602,
      "rewards/repetition_penalty_reward": -0.0517019834369421,
      "rewards/tag_count_reward": 0.6432291865348816,
      "step": 231
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 395.9271011352539,
      "epoch": 0.3479565054368204,
      "grad_norm": 4.9562590681699294,
      "kl": 0.408203125,
      "learning_rate": 1.6483506393303807e-05,
      "loss": 0.6441,
      "reward": 1.9554131031036377,
      "reward_std": 0.9711915105581284,
      "rewards/accuracy_reward": 0.484375,
      "rewards/reasoning_steps_reward": 0.7968750149011612,
      "rewards/repetition_penalty_reward": -0.04458692017942667,
      "rewards/tag_count_reward": 0.7187500298023224,
      "step": 232
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 427.2552261352539,
      "epoch": 0.34945631796025495,
      "grad_norm": 7.911974161092371,
      "kl": 0.58642578125,
      "learning_rate": 1.644348708726263e-05,
      "loss": 0.5395,
      "reward": 1.9124858379364014,
      "reward_std": 1.0289543271064758,
      "rewards/accuracy_reward": 0.5052083507180214,
      "rewards/reasoning_steps_reward": 0.7673611491918564,
      "rewards/repetition_penalty_reward": -0.046281606424599886,
      "rewards/tag_count_reward": 0.6861979365348816,
      "step": 233
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 377.78126525878906,
      "epoch": 0.35095613048368957,
      "grad_norm": 10.19550375864435,
      "kl": 0.98681640625,
      "learning_rate": 1.640329053945585e-05,
      "loss": 0.5746,
      "reward": 1.934269905090332,
      "reward_std": 1.0438069850206375,
      "rewards/accuracy_reward": 0.4687500074505806,
      "rewards/reasoning_steps_reward": 0.8072916865348816,
      "rewards/repetition_penalty_reward": -0.08526142686605453,
      "rewards/tag_count_reward": 0.743489608168602,
      "step": 234
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 407.5208511352539,
      "epoch": 0.35245594300712413,
      "grad_norm": 8.119699899556887,
      "kl": 0.48193359375,
      "learning_rate": 1.6362917855574694e-05,
      "loss": 0.8051,
      "reward": 1.804307907819748,
      "reward_std": 1.5109181106090546,
      "rewards/accuracy_reward": 0.5989583432674408,
      "rewards/reasoning_steps_reward": 0.7447916865348816,
      "rewards/repetition_penalty_reward": -0.24647344648838043,
      "rewards/tag_count_reward": 0.7070312649011612,
      "step": 235
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 353.2135543823242,
      "epoch": 0.3539557555305587,
      "grad_norm": 11.968853066603838,
      "kl": 1.50634765625,
      "learning_rate": 1.6322370146155372e-05,
      "loss": 0.6385,
      "reward": 1.9699607491493225,
      "reward_std": 1.293946921825409,
      "rewards/accuracy_reward": 0.5729166865348816,
      "rewards/reasoning_steps_reward": 0.8072916865348816,
      "rewards/repetition_penalty_reward": -0.19279972463846207,
      "rewards/tag_count_reward": 0.782552108168602,
      "step": 236
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 344.5416793823242,
      "epoch": 0.35545556805399325,
      "grad_norm": 10.830086232933182,
      "kl": 0.59326171875,
      "learning_rate": 1.6281648526548556e-05,
      "loss": 0.5964,
      "reward": 2.065182775259018,
      "reward_std": 1.1769357174634933,
      "rewards/accuracy_reward": 0.5781250298023224,
      "rewards/reasoning_steps_reward": 0.8420139104127884,
      "rewards/repetition_penalty_reward": -0.17266453802585602,
      "rewards/tag_count_reward": 0.8177083432674408,
      "step": 237
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 384.87500762939453,
      "epoch": 0.3569553805774278,
      "grad_norm": 13.320898103061964,
      "kl": 0.56640625,
      "learning_rate": 1.6240754116888673e-05,
      "loss": 0.5948,
      "reward": 1.9682828783988953,
      "reward_std": 1.2442002594470978,
      "rewards/accuracy_reward": 0.5416666865348816,
      "rewards/reasoning_steps_reward": 0.8263888955116272,
      "rewards/repetition_penalty_reward": -0.20315821841359138,
      "rewards/tag_count_reward": 0.8033854216337204,
      "step": 238
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 372.1354217529297,
      "epoch": 0.3584551931008624,
      "grad_norm": 12.247593958285009,
      "kl": 0.85302734375,
      "learning_rate": 1.6199688042063118e-05,
      "loss": 0.6045,
      "reward": 1.9835784435272217,
      "reward_std": 1.2273003607988358,
      "rewards/accuracy_reward": 0.5416666865348816,
      "rewards/reasoning_steps_reward": 0.8263889253139496,
      "rewards/repetition_penalty_reward": -0.1995813064277172,
      "rewards/tag_count_reward": 0.8151041865348816,
      "step": 239
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 385.89583587646484,
      "epoch": 0.35995500562429694,
      "grad_norm": 8.226772677936836,
      "kl": 0.62060546875,
      "learning_rate": 1.6158451431681292e-05,
      "loss": 0.5773,
      "reward": 1.98191300034523,
      "reward_std": 1.175494760274887,
      "rewards/accuracy_reward": 0.526041679084301,
      "rewards/reasoning_steps_reward": 0.8437500149011612,
      "rewards/repetition_penalty_reward": -0.19256622344255447,
      "rewards/tag_count_reward": 0.8046875149011612,
      "step": 240
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 338.96876525878906,
      "epoch": 0.36145481814773156,
      "grad_norm": 2.961510864903608,
      "kl": 0.4189453125,
      "learning_rate": 1.6117045420043545e-05,
      "loss": 0.4682,
      "reward": 2.1378689408302307,
      "reward_std": 0.9543499946594238,
      "rewards/accuracy_reward": 0.5625000149011612,
      "rewards/reasoning_steps_reward": 0.8784722238779068,
      "rewards/repetition_penalty_reward": -0.1546658743172884,
      "rewards/tag_count_reward": 0.8515625149011612,
      "step": 241
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 386.7760467529297,
      "epoch": 0.3629546306711661,
      "grad_norm": 4.568175543895563,
      "kl": 0.7275390625,
      "learning_rate": 1.6075471146109957e-05,
      "loss": 0.5672,
      "reward": 1.996113270521164,
      "reward_std": 0.9025778025388718,
      "rewards/accuracy_reward": 0.5468750149011612,
      "rewards/reasoning_steps_reward": 0.7881944626569748,
      "rewards/repetition_penalty_reward": -0.13192503154277802,
      "rewards/tag_count_reward": 0.7929687798023224,
      "step": 242
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 317.5260543823242,
      "epoch": 0.3644544431946007,
      "grad_norm": 6.725239225005341,
      "kl": 0.38720703125,
      "learning_rate": 1.603372975346903e-05,
      "loss": 0.4668,
      "reward": 2.1939920783042908,
      "reward_std": 0.7260987460613251,
      "rewards/accuracy_reward": 0.4635416865348816,
      "rewards/reasoning_steps_reward": 0.9236111342906952,
      "rewards/repetition_penalty_reward": -0.06946289446204901,
      "rewards/tag_count_reward": 0.8763020932674408,
      "step": 243
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 336.2552185058594,
      "epoch": 0.36595425571803525,
      "grad_norm": 6.395489506905219,
      "kl": 0.546875,
      "learning_rate": 1.599182239030621e-05,
      "loss": 0.5301,
      "reward": 2.0523171722888947,
      "reward_std": 0.8000431656837463,
      "rewards/accuracy_reward": 0.3958333432674408,
      "rewards/reasoning_steps_reward": 0.8888888955116272,
      "rewards/repetition_penalty_reward": -0.07094682566821575,
      "rewards/tag_count_reward": 0.8385417014360428,
      "step": 244
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 336.4635543823242,
      "epoch": 0.3674540682414698,
      "grad_norm": 1.7231966906953229,
      "kl": 0.44287109375,
      "learning_rate": 1.594975020937233e-05,
      "loss": 0.4191,
      "reward": 2.3042526245117188,
      "reward_std": 0.7582926452159882,
      "rewards/accuracy_reward": 0.6562500149011612,
      "rewards/reasoning_steps_reward": 0.8576389104127884,
      "rewards/repetition_penalty_reward": -0.07031344994902611,
      "rewards/tag_count_reward": 0.860677108168602,
      "step": 245
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 297.7083435058594,
      "epoch": 0.3689538807649044,
      "grad_norm": 78.49085937534032,
      "kl": 0.81298828125,
      "learning_rate": 1.590751436795186e-05,
      "loss": 0.5026,
      "reward": 2.197529286146164,
      "reward_std": 0.8267102539539337,
      "rewards/accuracy_reward": 0.4895833358168602,
      "rewards/reasoning_steps_reward": 0.8923611342906952,
      "rewards/repetition_penalty_reward": -0.07373820524662733,
      "rewards/tag_count_reward": 0.8893229365348816,
      "step": 246
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 352.8229293823242,
      "epoch": 0.37045369328833894,
      "grad_norm": 12.08436983860603,
      "kl": 0.72998046875,
      "learning_rate": 1.5865116027831123e-05,
      "loss": 0.6898,
      "reward": 2.0662292540073395,
      "reward_std": 1.1430340111255646,
      "rewards/accuracy_reward": 0.5312500074505806,
      "rewards/reasoning_steps_reward": 0.8368055671453476,
      "rewards/repetition_penalty_reward": -0.1273471899330616,
      "rewards/tag_count_reward": 0.825520858168602,
      "step": 247
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 339.5208435058594,
      "epoch": 0.37195350581177355,
      "grad_norm": 2.931100861119855,
      "kl": 0.4052734375,
      "learning_rate": 1.5822556355266302e-05,
      "loss": 0.5591,
      "reward": 2.0680652260780334,
      "reward_std": 1.04281547665596,
      "rewards/accuracy_reward": 0.5208333507180214,
      "rewards/reasoning_steps_reward": 0.8576388955116272,
      "rewards/repetition_penalty_reward": -0.15025084279477596,
      "rewards/tag_count_reward": 0.8398437649011612,
      "step": 248
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 367.25000762939453,
      "epoch": 0.3734533183352081,
      "grad_norm": 661.2695289482906,
      "kl": 5.46142578125,
      "learning_rate": 1.577983652095137e-05,
      "loss": 0.7068,
      "reward": 1.943572849035263,
      "reward_std": 1.0748744010925293,
      "rewards/accuracy_reward": 0.4270833507180214,
      "rewards/reasoning_steps_reward": 0.861111119389534,
      "rewards/repetition_penalty_reward": -0.1805591806769371,
      "rewards/tag_count_reward": 0.8359375298023224,
      "step": 249
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 305.01043701171875,
      "epoch": 0.3749531308586427,
      "grad_norm": 3.771451390000111,
      "kl": 0.345703125,
      "learning_rate": 1.5736957699985887e-05,
      "loss": 0.4327,
      "reward": 2.2907695174217224,
      "reward_std": 0.8615385890007019,
      "rewards/accuracy_reward": 0.6041666865348816,
      "rewards/reasoning_steps_reward": 0.927083358168602,
      "rewards/repetition_penalty_reward": -0.13761597499251366,
      "rewards/tag_count_reward": 0.8971354365348816,
      "step": 250
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 384.3489761352539,
      "epoch": 0.37645294338207724,
      "grad_norm": 194.96197462696594,
      "kl": 0.50048828125,
      "learning_rate": 1.5693921071842688e-05,
      "loss": 0.6103,
      "reward": 2.0306463837623596,
      "reward_std": 1.2375063300132751,
      "rewards/accuracy_reward": 0.5625000149011612,
      "rewards/reasoning_steps_reward": 0.850694477558136,
      "rewards/repetition_penalty_reward": -0.20155855640769005,
      "rewards/tag_count_reward": 0.8190104216337204,
      "step": 251
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 343.0833511352539,
      "epoch": 0.3779527559055118,
      "grad_norm": 14.689129884295669,
      "kl": 0.41357421875,
      "learning_rate": 1.5650727820335417e-05,
      "loss": 0.6158,
      "reward": 1.993709921836853,
      "reward_std": 1.1109019815921783,
      "rewards/accuracy_reward": 0.5104166716337204,
      "rewards/reasoning_steps_reward": 0.831597238779068,
      "rewards/repetition_penalty_reward": -0.18554353341460228,
      "rewards/tag_count_reward": 0.837239608168602,
      "step": 252
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 388.89064025878906,
      "epoch": 0.37945256842894637,
      "grad_norm": 14.608873859558672,
      "kl": 0.4140625,
      "learning_rate": 1.5607379133585978e-05,
      "loss": 0.7007,
      "reward": 2.0233902037143707,
      "reward_std": 1.345583826303482,
      "rewards/accuracy_reward": 0.6093750149011612,
      "rewards/reasoning_steps_reward": 0.835069477558136,
      "rewards/repetition_penalty_reward": -0.21662724763154984,
      "rewards/tag_count_reward": 0.7955729365348816,
      "step": 253
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 367.6927261352539,
      "epoch": 0.38095238095238093,
      "grad_norm": 10.596812649365775,
      "kl": 0.3828125,
      "learning_rate": 1.5563876203991856e-05,
      "loss": 0.6745,
      "reward": 1.9966700077056885,
      "reward_std": 1.2292464971542358,
      "rewards/accuracy_reward": 0.5208333507180214,
      "rewards/reasoning_steps_reward": 0.8454861640930176,
      "rewards/repetition_penalty_reward": -0.18735776841640472,
      "rewards/tag_count_reward": 0.8177083432674408,
      "step": 254
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 349.64064025878906,
      "epoch": 0.38245219347581555,
      "grad_norm": 6.633191911603578,
      "kl": 0.36279296875,
      "learning_rate": 1.55202202281933e-05,
      "loss": 0.5756,
      "reward": 2.1044811606407166,
      "reward_std": 1.2099826782941818,
      "rewards/accuracy_reward": 0.6041666939854622,
      "rewards/reasoning_steps_reward": 0.8472222536802292,
      "rewards/repetition_penalty_reward": -0.1776370257139206,
      "rewards/tag_count_reward": 0.8307291865348816,
      "step": 255
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 411.07813262939453,
      "epoch": 0.3839520059992501,
      "grad_norm": 1.8892823585301353,
      "kl": 0.4150390625,
      "learning_rate": 1.5476412407040445e-05,
      "loss": 0.7152,
      "reward": 1.811952918767929,
      "reward_std": 1.4293029010295868,
      "rewards/accuracy_reward": 0.572916679084301,
      "rewards/reasoning_steps_reward": 0.7361111491918564,
      "rewards/repetition_penalty_reward": -0.2223353162407875,
      "rewards/tag_count_reward": 0.7252604365348816,
      "step": 256
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 413.4166793823242,
      "epoch": 0.3854518185226847,
      "grad_norm": 2.9450172922219524,
      "kl": 0.57177734375,
      "learning_rate": 1.5432453945560223e-05,
      "loss": 0.6283,
      "reward": 1.983522891998291,
      "reward_std": 1.2600146383047104,
      "rewards/accuracy_reward": 0.5833333507180214,
      "rewards/reasoning_steps_reward": 0.7777778208255768,
      "rewards/repetition_penalty_reward": -0.13019242137670517,
      "rewards/tag_count_reward": 0.7526041865348816,
      "step": 257
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 445.9739685058594,
      "epoch": 0.38695163104611924,
      "grad_norm": 4.635567397393551,
      "kl": 0.43603515625,
      "learning_rate": 1.5388346052923268e-05,
      "loss": 0.5593,
      "reward": 2.0053739845752716,
      "reward_std": 1.152603343129158,
      "rewards/accuracy_reward": 0.5677083507180214,
      "rewards/reasoning_steps_reward": 0.7847222536802292,
      "rewards/repetition_penalty_reward": -0.0866400208324194,
      "rewards/tag_count_reward": 0.7395833432674408,
      "step": 258
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 511.49481201171875,
      "epoch": 0.3884514435695538,
      "grad_norm": 3.270106987863464,
      "kl": 0.4833984375,
      "learning_rate": 1.534408994241063e-05,
      "loss": 0.4957,
      "reward": 1.8226185142993927,
      "reward_std": 1.1505940705537796,
      "rewards/accuracy_reward": 0.447916679084301,
      "rewards/reasoning_steps_reward": 0.7465278059244156,
      "rewards/repetition_penalty_reward": -0.07495102658867836,
      "rewards/tag_count_reward": 0.7031250149011612,
      "step": 259
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 485.95314025878906,
      "epoch": 0.38995125609298836,
      "grad_norm": 2.75945020139557,
      "kl": 0.47900390625,
      "learning_rate": 1.5299686831380395e-05,
      "loss": 0.4981,
      "reward": 1.8642282485961914,
      "reward_std": 1.083889901638031,
      "rewards/accuracy_reward": 0.4427083432674408,
      "rewards/reasoning_steps_reward": 0.7708333730697632,
      "rewards/repetition_penalty_reward": -0.08108432777225971,
      "rewards/tag_count_reward": 0.731770858168602,
      "step": 260
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 483.5677185058594,
      "epoch": 0.3914510686164229,
      "grad_norm": 2.796799480256457,
      "kl": 1.189453125,
      "learning_rate": 1.5255137941234228e-05,
      "loss": 0.3239,
      "reward": 2.0941615104675293,
      "reward_std": 1.0367512106895447,
      "rewards/accuracy_reward": 0.5625000149011612,
      "rewards/reasoning_steps_reward": 0.817708358168602,
      "rewards/repetition_penalty_reward": -0.08292189985513687,
      "rewards/tag_count_reward": 0.796875,
      "step": 261
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 490.7343978881836,
      "epoch": 0.39295088113985754,
      "grad_norm": 1.775868555651766,
      "kl": 0.47119140625,
      "learning_rate": 1.5210444497383745e-05,
      "loss": 0.4116,
      "reward": 2.075753331184387,
      "reward_std": 1.1199206858873367,
      "rewards/accuracy_reward": 0.5625000074505806,
      "rewards/reasoning_steps_reward": 0.8177083432674408,
      "rewards/repetition_penalty_reward": -0.12216338887810707,
      "rewards/tag_count_reward": 0.8177083432674408,
      "step": 262
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 496.3958511352539,
      "epoch": 0.3944506936632921,
      "grad_norm": 33.57969017695382,
      "kl": 2.5966796875,
      "learning_rate": 1.5165607729216822e-05,
      "loss": 0.4809,
      "reward": 1.9561425149440765,
      "reward_std": 1.0842882692813873,
      "rewards/accuracy_reward": 0.432291679084301,
      "rewards/reasoning_steps_reward": 0.8541666865348816,
      "rewards/repetition_penalty_reward": -0.15453467145562172,
      "rewards/tag_count_reward": 0.8242187649011612,
      "step": 263
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 445.7448043823242,
      "epoch": 0.39595050618672667,
      "grad_norm": 5.931757796863058,
      "kl": 15.0380859375,
      "learning_rate": 1.5120628870063772e-05,
      "loss": 0.285,
      "reward": 2.2481858134269714,
      "reward_std": 0.9314106553792953,
      "rewards/accuracy_reward": 0.6145833507180214,
      "rewards/reasoning_steps_reward": 0.9184027910232544,
      "rewards/repetition_penalty_reward": -0.13115456514060497,
      "rewards/tag_count_reward": 0.8463541865348816,
      "step": 264
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 430.67188262939453,
      "epoch": 0.39745031871016123,
      "grad_norm": 4.7870139797358275,
      "kl": 0.74755859375,
      "learning_rate": 1.5075509157163422e-05,
      "loss": 0.3762,
      "reward": 2.189552366733551,
      "reward_std": 1.0156923830509186,
      "rewards/accuracy_reward": 0.6145833432674408,
      "rewards/reasoning_steps_reward": 0.897569477558136,
      "rewards/repetition_penalty_reward": -0.14942347817122936,
      "rewards/tag_count_reward": 0.8268229365348816,
      "step": 265
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 438.3958435058594,
      "epoch": 0.3989501312335958,
      "grad_norm": 7.646687103422767,
      "kl": 0.68505859375,
      "learning_rate": 1.503024983162908e-05,
      "loss": 0.4374,
      "reward": 2.135951668024063,
      "reward_std": 1.0654807686805725,
      "rewards/accuracy_reward": 0.5937500149011612,
      "rewards/reasoning_steps_reward": 0.8906250149011612,
      "rewards/repetition_penalty_reward": -0.17654842138290405,
      "rewards/tag_count_reward": 0.8281250149011612,
      "step": 266
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 425.43751525878906,
      "epoch": 0.40044994375703036,
      "grad_norm": 6.4245183766465415,
      "kl": 0.4541015625,
      "learning_rate": 1.49848521384144e-05,
      "loss": 0.4737,
      "reward": 2.125093847513199,
      "reward_std": 1.0813192874193192,
      "rewards/accuracy_reward": 0.5677083507180214,
      "rewards/reasoning_steps_reward": 0.8906250298023224,
      "rewards/repetition_penalty_reward": -0.16396865621209145,
      "rewards/tag_count_reward": 0.8307292014360428,
      "step": 267
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 344.36458587646484,
      "epoch": 0.4019497562804649,
      "grad_norm": 0.963320833591549,
      "kl": 0.65771484375,
      "learning_rate": 1.4939317326279125e-05,
      "loss": 0.3726,
      "reward": 2.1674315333366394,
      "reward_std": 0.8946598768234253,
      "rewards/accuracy_reward": 0.5781250074505806,
      "rewards/reasoning_steps_reward": 0.8697916865348816,
      "rewards/repetition_penalty_reward": -0.12423517927527428,
      "rewards/tag_count_reward": 0.8437500298023224,
      "step": 268
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 435.85938262939453,
      "epoch": 0.40344956880389954,
      "grad_norm": 0.7554295329920303,
      "kl": 0.61767578125,
      "learning_rate": 1.489364664775475e-05,
      "loss": 0.3679,
      "reward": 1.9763840436935425,
      "reward_std": 0.8836520612239838,
      "rewards/accuracy_reward": 0.588541679084301,
      "rewards/reasoning_steps_reward": 0.8715278059244156,
      "rewards/repetition_penalty_reward": -0.14514389261603355,
      "rewards/tag_count_reward": 0.6614583432674408,
      "step": 269
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 461.4635543823242,
      "epoch": 0.4049493813273341,
      "grad_norm": 1.4465486237113498,
      "kl": 1.75048828125,
      "learning_rate": 1.4847841359110058e-05,
      "loss": 0.2759,
      "reward": 2.053072690963745,
      "reward_std": 0.9807803928852081,
      "rewards/accuracy_reward": 0.6562500149011612,
      "rewards/reasoning_steps_reward": 0.8576388955116272,
      "rewards/repetition_penalty_reward": -0.1131600309163332,
      "rewards/tag_count_reward": 0.6523437649011612,
      "step": 270
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 526.7500152587891,
      "epoch": 0.40644919385076866,
      "grad_norm": 3.0949957954601928,
      "kl": 0.60693359375,
      "learning_rate": 1.480190272031657e-05,
      "loss": 0.3736,
      "reward": 1.8765326142311096,
      "reward_std": 1.0632285475730896,
      "rewards/accuracy_reward": 0.4739583469927311,
      "rewards/reasoning_steps_reward": 0.8211805522441864,
      "rewards/repetition_penalty_reward": -0.1126167606562376,
      "rewards/tag_count_reward": 0.6940104216337204,
      "step": 271
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 598.4635772705078,
      "epoch": 0.4079490063742032,
      "grad_norm": 2.8462767374545037,
      "kl": 0.43017578125,
      "learning_rate": 1.475583199501389e-05,
      "loss": 0.2644,
      "reward": 1.9981749355793,
      "reward_std": 1.0920402854681015,
      "rewards/accuracy_reward": 0.567708358168602,
      "rewards/reasoning_steps_reward": 0.8072916865348816,
      "rewards/repetition_penalty_reward": -0.13463760912418365,
      "rewards/tag_count_reward": 0.7578125298023224,
      "step": 272
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 624.3802337646484,
      "epoch": 0.4094488188976378,
      "grad_norm": 1.682946550152949,
      "kl": 0.3154296875,
      "learning_rate": 1.4709630450474936e-05,
      "loss": 0.3079,
      "reward": 1.9099436402320862,
      "reward_std": 1.3163824081420898,
      "rewards/accuracy_reward": 0.5989583507180214,
      "rewards/reasoning_steps_reward": 0.7760416865348816,
      "rewards/repetition_penalty_reward": -0.1812022142112255,
      "rewards/tag_count_reward": 0.716145858168602,
      "step": 273
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 662.1250152587891,
      "epoch": 0.41094863142107235,
      "grad_norm": 0.8488180233484801,
      "kl": 0.421875,
      "learning_rate": 1.466329935757109e-05,
      "loss": 0.2648,
      "reward": 1.6805398762226105,
      "reward_std": 1.3386418521404266,
      "rewards/accuracy_reward": 0.4687500074505806,
      "rewards/reasoning_steps_reward": 0.7447916865348816,
      "rewards/repetition_penalty_reward": -0.24263722822070122,
      "rewards/tag_count_reward": 0.7096354365348816,
      "step": 274
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 623.0469055175781,
      "epoch": 0.4124484439445069,
      "grad_norm": 3.2203198278586647,
      "kl": 0.39111328125,
      "learning_rate": 1.4616839990737232e-05,
      "loss": 0.2989,
      "reward": 1.7977931797504425,
      "reward_std": 1.3433575332164764,
      "rewards/accuracy_reward": 0.4739583358168602,
      "rewards/reasoning_steps_reward": 0.7916666865348816,
      "rewards/repetition_penalty_reward": -0.22955061122775078,
      "rewards/tag_count_reward": 0.7617187649011612,
      "step": 275
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 542.328125,
      "epoch": 0.41394825646794153,
      "grad_norm": 1.263395821949462,
      "kl": 0.92236328125,
      "learning_rate": 1.4570253627936693e-05,
      "loss": 0.2418,
      "reward": 2.0932159423828125,
      "reward_std": 1.0350356549024582,
      "rewards/accuracy_reward": 0.5416666865348816,
      "rewards/reasoning_steps_reward": 0.8593750298023224,
      "rewards/repetition_penalty_reward": -0.168502826243639,
      "rewards/tag_count_reward": 0.860677108168602,
      "step": 276
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 587.0781402587891,
      "epoch": 0.4154480689913761,
      "grad_norm": 0.6161220934962688,
      "kl": 0.4345703125,
      "learning_rate": 1.4523541550626093e-05,
      "loss": 0.1862,
      "reward": 2.1349116563796997,
      "reward_std": 0.9175398647785187,
      "rewards/accuracy_reward": 0.5677083432674408,
      "rewards/reasoning_steps_reward": 0.8854166865348816,
      "rewards/repetition_penalty_reward": -0.18409884721040726,
      "rewards/tag_count_reward": 0.8658854365348816,
      "step": 277
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 481.4427185058594,
      "epoch": 0.41694788151481066,
      "grad_norm": 0.6544307412376995,
      "kl": 1.18701171875,
      "learning_rate": 1.4476705043720099e-05,
      "loss": 0.1448,
      "reward": 2.3552486300468445,
      "reward_std": 0.7834517806768417,
      "rewards/accuracy_reward": 0.6614583507180214,
      "rewards/reasoning_steps_reward": 0.927083358168602,
      "rewards/repetition_penalty_reward": -0.1252202671021223,
      "rewards/tag_count_reward": 0.8919270932674408,
      "step": 278
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 469.12501525878906,
      "epoch": 0.4184476940382452,
      "grad_norm": 1.6940544921635285,
      "kl": 0.65087890625,
      "learning_rate": 1.4429745395556073e-05,
      "loss": 0.1976,
      "reward": 2.295401096343994,
      "reward_std": 0.7886304408311844,
      "rewards/accuracy_reward": 0.6145833432674408,
      "rewards/reasoning_steps_reward": 0.9079861342906952,
      "rewards/repetition_penalty_reward": -0.12300180085003376,
      "rewards/tag_count_reward": 0.8958333432674408,
      "step": 279
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 444.00521087646484,
      "epoch": 0.4199475065616798,
      "grad_norm": 2.02027175437092,
      "kl": 0.3291015625,
      "learning_rate": 1.4382663897858647e-05,
      "loss": 0.0773,
      "reward": 2.3556065559387207,
      "reward_std": 0.5934457406401634,
      "rewards/accuracy_reward": 0.5625000149011612,
      "rewards/reasoning_steps_reward": 0.9618055820465088,
      "rewards/repetition_penalty_reward": -0.10489701479673386,
      "rewards/tag_count_reward": 0.9361979365348816,
      "step": 280
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 408.82814025878906,
      "epoch": 0.42144731908511435,
      "grad_norm": 2.1520369019481316,
      "kl": 0.3623046875,
      "learning_rate": 1.4335461845704173e-05,
      "loss": 0.3092,
      "reward": 2.3092455863952637,
      "reward_std": 0.8222566097974777,
      "rewards/accuracy_reward": 0.6197916865348816,
      "rewards/reasoning_steps_reward": 0.9149305820465088,
      "rewards/repetition_penalty_reward": -0.1161017045378685,
      "rewards/tag_count_reward": 0.8906250298023224,
      "step": 281
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 389.1302185058594,
      "epoch": 0.4229471316085489,
      "grad_norm": 2.1074908824998912,
      "kl": 0.31689453125,
      "learning_rate": 1.428814053748512e-05,
      "loss": 0.2336,
      "reward": 2.3464564085006714,
      "reward_std": 0.8324245363473892,
      "rewards/accuracy_reward": 0.6666666716337204,
      "rewards/reasoning_steps_reward": 0.9166666865348816,
      "rewards/repetition_penalty_reward": -0.13140839524567127,
      "rewards/tag_count_reward": 0.8945312798023224,
      "step": 282
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 382.51564025878906,
      "epoch": 0.42444694413198353,
      "grad_norm": 2.4209991865943645,
      "kl": 0.30810546875,
      "learning_rate": 1.4240701274874331e-05,
      "loss": 0.3742,
      "reward": 2.1715636253356934,
      "reward_std": 0.9031594395637512,
      "rewards/accuracy_reward": 0.5729166865348816,
      "rewards/reasoning_steps_reward": 0.8697916865348816,
      "rewards/repetition_penalty_reward": -0.1435406319797039,
      "rewards/tag_count_reward": 0.872395858168602,
      "step": 283
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 351.92188262939453,
      "epoch": 0.4259467566554181,
      "grad_norm": 2.6366018633296884,
      "kl": 0.48486328125,
      "learning_rate": 1.419314536278925e-05,
      "loss": 0.5396,
      "reward": 2.206935942173004,
      "reward_std": 1.114256203174591,
      "rewards/accuracy_reward": 0.6510416865348816,
      "rewards/reasoning_steps_reward": 0.8559028208255768,
      "rewards/repetition_penalty_reward": -0.1619878150522709,
      "rewards/tag_count_reward": 0.8619791865348816,
      "step": 284
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 305.18751525878906,
      "epoch": 0.42744656917885265,
      "grad_norm": 1.5196765033206825,
      "kl": 0.53564453125,
      "learning_rate": 1.4145474109356008e-05,
      "loss": 0.2239,
      "reward": 2.3387808799743652,
      "reward_std": 0.7637557685375214,
      "rewards/accuracy_reward": 0.6458333432674408,
      "rewards/reasoning_steps_reward": 0.9131944626569748,
      "rewards/repetition_penalty_reward": -0.10696578957140446,
      "rewards/tag_count_reward": 0.8867187649011612,
      "step": 285
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 367.5989685058594,
      "epoch": 0.4289463817022872,
      "grad_norm": 1.1289332268594263,
      "kl": 0.37158203125,
      "learning_rate": 1.4097688825873437e-05,
      "loss": 0.3786,
      "reward": 2.217616856098175,
      "reward_std": 0.9250410199165344,
      "rewards/accuracy_reward": 0.6354166865348816,
      "rewards/reasoning_steps_reward": 0.897569477558136,
      "rewards/repetition_penalty_reward": -0.1643276885151863,
      "rewards/tag_count_reward": 0.848958358168602,
      "step": 286
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 383.53126525878906,
      "epoch": 0.4304461942257218,
      "grad_norm": 0.8172511360961945,
      "kl": 0.3134765625,
      "learning_rate": 1.4049790826777016e-05,
      "loss": 0.316,
      "reward": 2.058010071516037,
      "reward_std": 0.7540641278028488,
      "rewards/accuracy_reward": 0.4270833432674408,
      "rewards/reasoning_steps_reward": 0.911458358168602,
      "rewards/repetition_penalty_reward": -0.13990668766200542,
      "rewards/tag_count_reward": 0.8593750298023224,
      "step": 287
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 366.4479293823242,
      "epoch": 0.43194600674915634,
      "grad_norm": 2.8494106639784276,
      "kl": 0.578125,
      "learning_rate": 1.4001781429602704e-05,
      "loss": 0.2977,
      "reward": 2.2450991570949554,
      "reward_std": 0.8115597367286682,
      "rewards/accuracy_reward": 0.5572916716337204,
      "rewards/reasoning_steps_reward": 0.9270833283662796,
      "rewards/repetition_penalty_reward": -0.11036965623497963,
      "rewards/tag_count_reward": 0.8710937649011612,
      "step": 288
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 362.0573043823242,
      "epoch": 0.4334458192725909,
      "grad_norm": 1.5771836903904428,
      "kl": 0.34228515625,
      "learning_rate": 1.3953661954950693e-05,
      "loss": 0.213,
      "reward": 2.2435672879219055,
      "reward_std": 0.7003582268953323,
      "rewards/accuracy_reward": 0.5260416865348816,
      "rewards/reasoning_steps_reward": 0.9461805820465088,
      "rewards/repetition_penalty_reward": -0.11407159268856049,
      "rewards/tag_count_reward": 0.8854166865348816,
      "step": 289
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 314.3020935058594,
      "epoch": 0.4349456317960255,
      "grad_norm": 1.307478764620523,
      "kl": 0.40576171875,
      "learning_rate": 1.3905433726449102e-05,
      "loss": 0.1718,
      "reward": 2.420071065425873,
      "reward_std": 0.5883182883262634,
      "rewards/accuracy_reward": 0.6041666865348816,
      "rewards/reasoning_steps_reward": 0.9652777761220932,
      "rewards/repetition_penalty_reward": -0.08426924794912338,
      "rewards/tag_count_reward": 0.934895858168602,
      "step": 290
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 344.4270935058594,
      "epoch": 0.4364454443194601,
      "grad_norm": 1.066920831840518,
      "kl": 0.69140625,
      "learning_rate": 1.3857098070717543e-05,
      "loss": 0.2395,
      "reward": 2.4870429635047913,
      "reward_std": 0.7747205495834351,
      "rewards/accuracy_reward": 0.7291666939854622,
      "rewards/reasoning_steps_reward": 0.9461805820465088,
      "rewards/repetition_penalty_reward": -0.10106483101844788,
      "rewards/tag_count_reward": 0.9127604365348816,
      "step": 291
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 349.6822967529297,
      "epoch": 0.43794525684289465,
      "grad_norm": 0.737418903498002,
      "kl": 0.33349609375,
      "learning_rate": 1.3808656317330646e-05,
      "loss": 0.1173,
      "reward": 2.453067660331726,
      "reward_std": 0.43977494165301323,
      "rewards/accuracy_reward": 0.6406250298023224,
      "rewards/reasoning_steps_reward": 0.9809028059244156,
      "rewards/repetition_penalty_reward": -0.10465812310576439,
      "rewards/tag_count_reward": 0.9361979365348816,
      "step": 292
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 384.39063262939453,
      "epoch": 0.4394450693663292,
      "grad_norm": 1.8989977209226134,
      "kl": 0.5810546875,
      "learning_rate": 1.3760109798781489e-05,
      "loss": 0.3604,
      "reward": 2.2495803833007812,
      "reward_std": 0.849450945854187,
      "rewards/accuracy_reward": 0.5625000149011612,
      "rewards/reasoning_steps_reward": 0.927083358168602,
      "rewards/repetition_penalty_reward": -0.13974259793758392,
      "rewards/tag_count_reward": 0.8997395932674408,
      "step": 293
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 369.5833435058594,
      "epoch": 0.4409448818897638,
      "grad_norm": 0.5939727210211808,
      "kl": 0.7841796875,
      "learning_rate": 1.3711459850444923e-05,
      "loss": 0.2594,
      "reward": 2.291065901517868,
      "reward_std": 0.7707606852054596,
      "rewards/accuracy_reward": 0.5416666716337204,
      "rewards/reasoning_steps_reward": 0.9479166865348816,
      "rewards/repetition_penalty_reward": -0.11648627929389477,
      "rewards/tag_count_reward": 0.9179687649011612,
      "step": 294
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 386.82813262939453,
      "epoch": 0.44244469441319834,
      "grad_norm": 2.0394609526475236,
      "kl": 0.73681640625,
      "learning_rate": 1.3662707810540867e-05,
      "loss": 0.3294,
      "reward": 2.276679277420044,
      "reward_std": 1.02875255048275,
      "rewards/accuracy_reward": 0.6354166865348816,
      "rewards/reasoning_steps_reward": 0.8888889104127884,
      "rewards/repetition_penalty_reward": -0.13825136795639992,
      "rewards/tag_count_reward": 0.8906250298023224,
      "step": 295
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 359.828125,
      "epoch": 0.4439445069366329,
      "grad_norm": 0.9901125024422729,
      "kl": 0.392578125,
      "learning_rate": 1.3613855020097477e-05,
      "loss": 0.2768,
      "reward": 2.5605881214141846,
      "reward_std": 0.7392320334911346,
      "rewards/accuracy_reward": 0.8177083432674408,
      "rewards/reasoning_steps_reward": 0.9392361491918564,
      "rewards/repetition_penalty_reward": -0.12083561439067125,
      "rewards/tag_count_reward": 0.9244791716337204,
      "step": 296
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 383.4427185058594,
      "epoch": 0.4454443194600675,
      "grad_norm": 0.6075657581629706,
      "kl": 1.2001953125,
      "learning_rate": 1.3564902822914274e-05,
      "loss": 0.3837,
      "reward": 2.383248746395111,
      "reward_std": 1.0118870586156845,
      "rewards/accuracy_reward": 0.7239583432674408,
      "rewards/reasoning_steps_reward": 0.907986119389534,
      "rewards/repetition_penalty_reward": -0.1510394662618637,
      "rewards/tag_count_reward": 0.9023437798023224,
      "step": 297
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 421.82813262939453,
      "epoch": 0.4469441319835021,
      "grad_norm": 0.7992325861558693,
      "kl": 0.39208984375,
      "learning_rate": 1.3515852565525167e-05,
      "loss": 0.4008,
      "reward": 2.303019016981125,
      "reward_std": 1.0580395609140396,
      "rewards/accuracy_reward": 0.723958358168602,
      "rewards/reasoning_steps_reward": 0.8854166716337204,
      "rewards/repetition_penalty_reward": -0.1839602366089821,
      "rewards/tag_count_reward": 0.8776042014360428,
      "step": 298
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 360.6666793823242,
      "epoch": 0.44844394450693664,
      "grad_norm": 0.7019719444995071,
      "kl": 0.45654296875,
      "learning_rate": 1.3466705597161416e-05,
      "loss": 0.0816,
      "reward": 2.5587180852890015,
      "reward_std": 0.4348931238055229,
      "rewards/accuracy_reward": 0.6927083656191826,
      "rewards/reasoning_steps_reward": 0.9791666716337204,
      "rewards/repetition_penalty_reward": -0.08971942402422428,
      "rewards/tag_count_reward": 0.9765625149011612,
      "step": 299
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 360.20833587646484,
      "epoch": 0.4499437570303712,
      "grad_norm": 0.8473895074586508,
      "kl": 0.51953125,
      "learning_rate": 1.3417463269714525e-05,
      "loss": 0.1846,
      "reward": 2.5228134989738464,
      "reward_std": 0.6405130326747894,
      "rewards/accuracy_reward": 0.6927083432674408,
      "rewards/reasoning_steps_reward": 0.9652777910232544,
      "rewards/repetition_penalty_reward": -0.09350610896945,
      "rewards/tag_count_reward": 0.9583333432674408,
      "step": 300
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 404.3698043823242,
      "epoch": 0.45144356955380577,
      "grad_norm": 1.59719244073744,
      "kl": 0.48876953125,
      "learning_rate": 1.3368126937699055e-05,
      "loss": 0.3579,
      "reward": 2.2757957577705383,
      "reward_std": 0.9483175575733185,
      "rewards/accuracy_reward": 0.6145833507180214,
      "rewards/reasoning_steps_reward": 0.904513880610466,
      "rewards/repetition_penalty_reward": -0.14694742858409882,
      "rewards/tag_count_reward": 0.903645858168602,
      "step": 301
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 378.29688262939453,
      "epoch": 0.45294338207724033,
      "grad_norm": 0.47093010825160103,
      "kl": 0.86376953125,
      "learning_rate": 1.3318697958215358e-05,
      "loss": 0.3148,
      "reward": 2.259345531463623,
      "reward_std": 1.0043332874774933,
      "rewards/accuracy_reward": 0.5885416716337204,
      "rewards/reasoning_steps_reward": 0.9062500298023224,
      "rewards/repetition_penalty_reward": -0.13258161395788193,
      "rewards/tag_count_reward": 0.8971354514360428,
      "step": 302
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 426.34376525878906,
      "epoch": 0.4544431946006749,
      "grad_norm": 1.6840109923746656,
      "kl": 0.615234375,
      "learning_rate": 1.3269177690912244e-05,
      "loss": 0.5099,
      "reward": 2.000718355178833,
      "reward_std": 1.0972090363502502,
      "rewards/accuracy_reward": 0.494791679084301,
      "rewards/reasoning_steps_reward": 0.8593750149011612,
      "rewards/repetition_penalty_reward": -0.20370880514383316,
      "rewards/tag_count_reward": 0.8502604514360428,
      "step": 303
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 503.51564025878906,
      "epoch": 0.4559430071241095,
      "grad_norm": 0.8957611247139153,
      "kl": 0.5615234375,
      "learning_rate": 1.3219567497949603e-05,
      "loss": 0.5607,
      "reward": 1.748295396566391,
      "reward_std": 1.3299023509025574,
      "rewards/accuracy_reward": 0.4270833469927311,
      "rewards/reasoning_steps_reward": 0.8368055522441864,
      "rewards/repetition_penalty_reward": -0.311166375875473,
      "rewards/tag_count_reward": 0.7955729514360428,
      "step": 304
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 463.82813262939453,
      "epoch": 0.4574428196475441,
      "grad_norm": 0.7521310552587273,
      "kl": 0.373046875,
      "learning_rate": 1.3169868743960904e-05,
      "loss": 0.5557,
      "reward": 2.0611188113689423,
      "reward_std": 1.065147504210472,
      "rewards/accuracy_reward": 0.5364583432674408,
      "rewards/reasoning_steps_reward": 0.9236111342906952,
      "rewards/repetition_penalty_reward": -0.25572141259908676,
      "rewards/tag_count_reward": 0.856770858168602,
      "step": 305
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 403.75001525878906,
      "epoch": 0.45894263217097864,
      "grad_norm": 0.531106231660718,
      "kl": 0.33935546875,
      "learning_rate": 1.3120082796015694e-05,
      "loss": 0.4341,
      "reward": 2.1974433958530426,
      "reward_std": 0.9661893397569656,
      "rewards/accuracy_reward": 0.5468750074505806,
      "rewards/reasoning_steps_reward": 0.9270833432674408,
      "rewards/repetition_penalty_reward": -0.17234836518764496,
      "rewards/tag_count_reward": 0.895833358168602,
      "step": 306
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 389.5104293823242,
      "epoch": 0.4604424446944132,
      "grad_norm": 0.9868194546337975,
      "kl": 8.66748046875,
      "learning_rate": 1.3070211023581959e-05,
      "loss": 0.5127,
      "reward": 2.096434473991394,
      "reward_std": 1.0376380234956741,
      "rewards/accuracy_reward": 0.5000000260770321,
      "rewards/reasoning_steps_reward": 0.9027778208255768,
      "rewards/repetition_penalty_reward": -0.18134328350424767,
      "rewards/tag_count_reward": 0.8750000149011612,
      "step": 307
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 346.2239761352539,
      "epoch": 0.46194225721784776,
      "grad_norm": 1.1602735630677965,
      "kl": 0.60205078125,
      "learning_rate": 1.302025479848847e-05,
      "loss": 0.6158,
      "reward": 2.2226256132125854,
      "reward_std": 1.0068841725587845,
      "rewards/accuracy_reward": 0.5625000149011612,
      "rewards/reasoning_steps_reward": 0.9236111491918564,
      "rewards/repetition_penalty_reward": -0.15801683440804482,
      "rewards/tag_count_reward": 0.8945312649011612,
      "step": 308
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 365.1145935058594,
      "epoch": 0.4634420697412823,
      "grad_norm": 0.5983467148707611,
      "kl": 0.783203125,
      "learning_rate": 1.2970215494887057e-05,
      "loss": 0.4409,
      "reward": 2.213726222515106,
      "reward_std": 0.9471542239189148,
      "rewards/accuracy_reward": 0.5312500149011612,
      "rewards/reasoning_steps_reward": 0.9375000298023224,
      "rewards/repetition_penalty_reward": -0.1339300237596035,
      "rewards/tag_count_reward": 0.8789062649011612,
      "step": 309
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 383.40626525878906,
      "epoch": 0.4649418822647169,
      "grad_norm": 0.5734600272497099,
      "kl": 0.890625,
      "learning_rate": 1.2920094489214794e-05,
      "loss": 0.4356,
      "reward": 2.073159396648407,
      "reward_std": 1.0565957874059677,
      "rewards/accuracy_reward": 0.4843750223517418,
      "rewards/reasoning_steps_reward": 0.9062500298023224,
      "rewards/repetition_penalty_reward": -0.16642405465245247,
      "rewards/tag_count_reward": 0.848958358168602,
      "step": 310
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 301.56250762939453,
      "epoch": 0.4664416947881515,
      "grad_norm": 0.7211277806322656,
      "kl": 0.6845703125,
      "learning_rate": 1.2869893160156144e-05,
      "loss": 0.2848,
      "reward": 2.1854459047317505,
      "reward_std": 0.8107419461011887,
      "rewards/accuracy_reward": 0.4375000149011612,
      "rewards/reasoning_steps_reward": 0.9322916865348816,
      "rewards/repetition_penalty_reward": -0.1036166287958622,
      "rewards/tag_count_reward": 0.9192708432674408,
      "step": 311
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 323.3333511352539,
      "epoch": 0.46794150731158607,
      "grad_norm": 1.8610423986745928,
      "kl": 0.54345703125,
      "learning_rate": 1.2819612888605038e-05,
      "loss": 0.4624,
      "reward": 2.25553822517395,
      "reward_std": 1.062050774693489,
      "rewards/accuracy_reward": 0.6406250149011612,
      "rewards/reasoning_steps_reward": 0.880208358168602,
      "rewards/repetition_penalty_reward": -0.13769106939435005,
      "rewards/tag_count_reward": 0.872395858168602,
      "step": 312
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 269.31771087646484,
      "epoch": 0.46944131983502063,
      "grad_norm": 1.106507460021578,
      "kl": 1.8193359375,
      "learning_rate": 1.2769255057626879e-05,
      "loss": 0.0841,
      "reward": 2.6005271077156067,
      "reward_std": 0.4192248545587063,
      "rewards/accuracy_reward": 0.7291666939854622,
      "rewards/reasoning_steps_reward": 0.9687500149011612,
      "rewards/repetition_penalty_reward": -0.06353538297116756,
      "rewards/tag_count_reward": 0.9661458432674408,
      "step": 313
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 262.1979217529297,
      "epoch": 0.4709411323584552,
      "grad_norm": 11.712890420404275,
      "kl": 12.896484375,
      "learning_rate": 1.2718821052420518e-05,
      "loss": 0.4463,
      "reward": 2.492148220539093,
      "reward_std": 0.6998696699738503,
      "rewards/accuracy_reward": 0.6822917014360428,
      "rewards/reasoning_steps_reward": 0.9461805671453476,
      "rewards/repetition_penalty_reward": -0.0842407438904047,
      "rewards/tag_count_reward": 0.9479167014360428,
      "step": 314
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 265.77605056762695,
      "epoch": 0.47244094488188976,
      "grad_norm": 0.6975411820310382,
      "kl": 1.18310546875,
      "learning_rate": 1.2668312260280136e-05,
      "loss": 0.1567,
      "reward": 2.5813058018684387,
      "reward_std": 0.6979668289422989,
      "rewards/accuracy_reward": 0.7656250149011612,
      "rewards/reasoning_steps_reward": 0.9375,
      "rewards/repetition_penalty_reward": -0.07103797234594822,
      "rewards/tag_count_reward": 0.9492187649011612,
      "step": 315
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 237.0989646911621,
      "epoch": 0.4739407574053243,
      "grad_norm": 0.9429145740895207,
      "kl": 0.6083984375,
      "learning_rate": 1.2617730070557079e-05,
      "loss": 0.1167,
      "reward": 2.5124951004981995,
      "reward_std": 0.5172496251761913,
      "rewards/accuracy_reward": 0.619791679084301,
      "rewards/reasoning_steps_reward": 0.9739583432674408,
      "rewards/repetition_penalty_reward": -0.05391131527721882,
      "rewards/tag_count_reward": 0.97265625,
      "step": 316
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 268.03125762939453,
      "epoch": 0.4754405699287589,
      "grad_norm": 0.8535133809055152,
      "kl": 0.49755859375,
      "learning_rate": 1.2567075874621658e-05,
      "loss": 0.1603,
      "reward": 2.632077991962433,
      "reward_std": 0.5518362149596214,
      "rewards/accuracy_reward": 0.7656250298023224,
      "rewards/reasoning_steps_reward": 0.9687500447034836,
      "rewards/repetition_penalty_reward": -0.06583872064948082,
      "rewards/tag_count_reward": 0.9635416716337204,
      "step": 317
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 311.51564025878906,
      "epoch": 0.4769403824521935,
      "grad_norm": 0.7780541688251532,
      "kl": 0.5283203125,
      "learning_rate": 1.2516351065824864e-05,
      "loss": 0.1901,
      "reward": 2.402199685573578,
      "reward_std": 0.6393345445394516,
      "rewards/accuracy_reward": 0.5989583432674408,
      "rewards/reasoning_steps_reward": 0.9548611491918564,
      "rewards/repetition_penalty_reward": -0.10083861276507378,
      "rewards/tag_count_reward": 0.9492187649011612,
      "step": 318
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 299.9947967529297,
      "epoch": 0.47844019497562806,
      "grad_norm": 0.7104797645236225,
      "kl": 0.8134765625,
      "learning_rate": 1.2465557039460048e-05,
      "loss": 0.3304,
      "reward": 2.40736585855484,
      "reward_std": 0.7196965366601944,
      "rewards/accuracy_reward": 0.6197916865348816,
      "rewards/reasoning_steps_reward": 0.94618059694767,
      "rewards/repetition_penalty_reward": -0.10001271404325962,
      "rewards/tag_count_reward": 0.9414062798023224,
      "step": 319
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 326.98439025878906,
      "epoch": 0.4799400074990626,
      "grad_norm": 0.7194211623846433,
      "kl": 0.42431640625,
      "learning_rate": 1.241469519272453e-05,
      "loss": 0.5949,
      "reward": 2.308800458908081,
      "reward_std": 0.9084680825471878,
      "rewards/accuracy_reward": 0.6093750149011612,
      "rewards/reasoning_steps_reward": 0.9427083432674408,
      "rewards/repetition_penalty_reward": -0.1495329923927784,
      "rewards/tag_count_reward": 0.9062500298023224,
      "step": 320
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 357.56771087646484,
      "epoch": 0.4814398200224972,
      "grad_norm": 0.9593534999944733,
      "kl": 0.48095703125,
      "learning_rate": 1.2363766924681178e-05,
      "loss": 0.5733,
      "reward": 2.4493626952171326,
      "reward_std": 0.9347756206989288,
      "rewards/accuracy_reward": 0.7447917014360428,
      "rewards/reasoning_steps_reward": 0.9496527910232544,
      "rewards/repetition_penalty_reward": -0.15914421156048775,
      "rewards/tag_count_reward": 0.9140625298023224,
      "step": 321
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 293.7448043823242,
      "epoch": 0.48293963254593175,
      "grad_norm": 4.081284787550109,
      "kl": 0.697265625,
      "learning_rate": 1.2312773636219919e-05,
      "loss": 0.5756,
      "reward": 2.2688207626342773,
      "reward_std": 0.8435943201184273,
      "rewards/accuracy_reward": 0.5364583507180214,
      "rewards/reasoning_steps_reward": 0.9288194924592972,
      "rewards/repetition_penalty_reward": -0.12614460662007332,
      "rewards/tag_count_reward": 0.9296875149011612,
      "step": 322
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 309.8489685058594,
      "epoch": 0.4844394450693663,
      "grad_norm": 1.404497176795523,
      "kl": 1.14306640625,
      "learning_rate": 1.2261716730019202e-05,
      "loss": 0.5032,
      "reward": 2.3895642161369324,
      "reward_std": 0.8716422617435455,
      "rewards/accuracy_reward": 0.6770833432674408,
      "rewards/reasoning_steps_reward": 0.9479167014360428,
      "rewards/repetition_penalty_reward": -0.14819632470607758,
      "rewards/tag_count_reward": 0.9127604216337204,
      "step": 323
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 249.86459732055664,
      "epoch": 0.4859392575928009,
      "grad_norm": 1.926446222656527,
      "kl": 0.4814453125,
      "learning_rate": 1.2210597610507418e-05,
      "loss": 0.4739,
      "reward": 2.306950032711029,
      "reward_std": 0.7142234891653061,
      "rewards/accuracy_reward": 0.510416679084301,
      "rewards/reasoning_steps_reward": 0.9548611491918564,
      "rewards/repetition_penalty_reward": -0.10884871147572994,
      "rewards/tag_count_reward": 0.950520858168602,
      "step": 324
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 254.2187614440918,
      "epoch": 0.4874390701162355,
      "grad_norm": 5.256123203383221,
      "kl": 0.7255859375,
      "learning_rate": 1.2159417683824266e-05,
      "loss": 0.5998,
      "reward": 2.4224319458007812,
      "reward_std": 0.787856787443161,
      "rewards/accuracy_reward": 0.6666666865348816,
      "rewards/reasoning_steps_reward": 0.9357639104127884,
      "rewards/repetition_penalty_reward": -0.12270701117813587,
      "rewards/tag_count_reward": 0.942708358168602,
      "step": 325
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 327.9114685058594,
      "epoch": 0.48893888263967006,
      "grad_norm": 9.312199613733055,
      "kl": 1.2685546875,
      "learning_rate": 1.2108178357782079e-05,
      "loss": 0.8976,
      "reward": 1.884236216545105,
      "reward_std": 1.2027931809425354,
      "rewards/accuracy_reward": 0.4375000074505806,
      "rewards/reasoning_steps_reward": 0.8020833730697632,
      "rewards/repetition_penalty_reward": -0.2108159288764,
      "rewards/tag_count_reward": 0.8554687649011612,
      "step": 326
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 275.9791717529297,
      "epoch": 0.4904386951631046,
      "grad_norm": 15.398650924864615,
      "kl": 0.73828125,
      "learning_rate": 1.205688104182709e-05,
      "loss": 0.7201,
      "reward": 2.0736686289310455,
      "reward_std": 0.9543730318546295,
      "rewards/accuracy_reward": 0.4635416939854622,
      "rewards/reasoning_steps_reward": 0.866319477558136,
      "rewards/repetition_penalty_reward": -0.1533280350267887,
      "rewards/tag_count_reward": 0.8971354365348816,
      "step": 327
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 241.10417556762695,
      "epoch": 0.4919385076865392,
      "grad_norm": 37.30122920300535,
      "kl": 0.7822265625,
      "learning_rate": 1.2005527147000663e-05,
      "loss": 0.7101,
      "reward": 2.245617926120758,
      "reward_std": 0.8258701711893082,
      "rewards/accuracy_reward": 0.552083358168602,
      "rewards/reasoning_steps_reward": 0.9027778059244156,
      "rewards/repetition_penalty_reward": -0.12070165574550629,
      "rewards/tag_count_reward": 0.911458358168602,
      "step": 328
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 215.0520896911621,
      "epoch": 0.49343832020997375,
      "grad_norm": 60.05024321951368,
      "kl": 0.625,
      "learning_rate": 1.1954118085900503e-05,
      "loss": 0.5475,
      "reward": 2.432074725627899,
      "reward_std": 0.7434251010417938,
      "rewards/accuracy_reward": 0.6354167014360428,
      "rewards/reasoning_steps_reward": 0.9479166865348816,
      "rewards/repetition_penalty_reward": -0.08224830403923988,
      "rewards/tag_count_reward": 0.9309895932674408,
      "step": 329
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 224.40625381469727,
      "epoch": 0.4949381327334083,
      "grad_norm": 98.14120807441263,
      "kl": 0.5615234375,
      "learning_rate": 1.1902655272641756e-05,
      "loss": 0.5814,
      "reward": 2.4700867533683777,
      "reward_std": 0.6907303482294083,
      "rewards/accuracy_reward": 0.6718750298023224,
      "rewards/reasoning_steps_reward": 0.9392361342906952,
      "rewards/repetition_penalty_reward": -0.08112853486090899,
      "rewards/tag_count_reward": 0.9401041865348816,
      "step": 330
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 227.70833587646484,
      "epoch": 0.49643794525684287,
      "grad_norm": 28.380941791399632,
      "kl": 0.796875,
      "learning_rate": 1.1851140122818155e-05,
      "loss": 0.6384,
      "reward": 2.440628468990326,
      "reward_std": 0.5516751855611801,
      "rewards/accuracy_reward": 0.5937500149011612,
      "rewards/reasoning_steps_reward": 0.9652777761220932,
      "rewards/repetition_penalty_reward": -0.05069108493626118,
      "rewards/tag_count_reward": 0.9322917014360428,
      "step": 331
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 189.9635467529297,
      "epoch": 0.4979377577802775,
      "grad_norm": 90.64584397678644,
      "kl": 0.7421875,
      "learning_rate": 1.1799574053463048e-05,
      "loss": 0.2312,
      "reward": 2.4951672554016113,
      "reward_std": 0.4933183267712593,
      "rewards/accuracy_reward": 0.6145833432674408,
      "rewards/reasoning_steps_reward": 0.9774305820465088,
      "rewards/repetition_penalty_reward": -0.04476337507367134,
      "rewards/tag_count_reward": 0.9479166865348816,
      "step": 332
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 198.1822967529297,
      "epoch": 0.49943757030371205,
      "grad_norm": 47.44537979910489,
      "kl": 0.79296875,
      "learning_rate": 1.1747958483010438e-05,
      "loss": 0.1878,
      "reward": 2.4803009629249573,
      "reward_std": 0.4909311309456825,
      "rewards/accuracy_reward": 0.614583358168602,
      "rewards/reasoning_steps_reward": 0.9687500447034836,
      "rewards/repetition_penalty_reward": -0.054855335503816605,
      "rewards/tag_count_reward": 0.9518229365348816,
      "step": 333
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 213.3489646911621,
      "epoch": 0.5009373828271466,
      "grad_norm": 320.06224533206546,
      "kl": 2.462890625,
      "learning_rate": 1.1696294831255961e-05,
      "loss": 0.9267,
      "reward": 2.6039064526557922,
      "reward_std": 0.5018587484955788,
      "rewards/accuracy_reward": 0.7500000149011612,
      "rewards/reasoning_steps_reward": 0.960069477558136,
      "rewards/repetition_penalty_reward": -0.04366318695247173,
      "rewards/tag_count_reward": 0.9375000298023224,
      "step": 334
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 173.62500381469727,
      "epoch": 0.5024371953505812,
      "grad_norm": 1405.8210110352265,
      "kl": 1.1259765625,
      "learning_rate": 1.1644584519317828e-05,
      "loss": 0.2665,
      "reward": 2.5448213815689087,
      "reward_std": 0.4410114288330078,
      "rewards/accuracy_reward": 0.6302083358168602,
      "rewards/reasoning_steps_reward": 0.9809027761220932,
      "rewards/repetition_penalty_reward": -0.03764389827847481,
      "rewards/tag_count_reward": 0.9713541865348816,
      "step": 335
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 197.0885467529297,
      "epoch": 0.5039370078740157,
      "grad_norm": 266.36597808183785,
      "kl": 1.8984375,
      "learning_rate": 1.159282896959774e-05,
      "loss": 0.6665,
      "reward": 2.542451798915863,
      "reward_std": 0.4349544197320938,
      "rewards/accuracy_reward": 0.6562500223517418,
      "rewards/reasoning_steps_reward": 0.9826388955116272,
      "rewards/repetition_penalty_reward": -0.04695798270404339,
      "rewards/tag_count_reward": 0.9505208432674408,
      "step": 336
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 203.6666717529297,
      "epoch": 0.5054368203974503,
      "grad_norm": 124.39297508584644,
      "kl": 1.185546875,
      "learning_rate": 1.1541029605741758e-05,
      "loss": 0.6041,
      "reward": 2.4829863905906677,
      "reward_std": 0.5544790178537369,
      "rewards/accuracy_reward": 0.6093750298023224,
      "rewards/reasoning_steps_reward": 0.9791666865348816,
      "rewards/repetition_penalty_reward": -0.043055362068116665,
      "rewards/tag_count_reward": 0.9375000298023224,
      "step": 337
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 182.04167556762695,
      "epoch": 0.5069366329208849,
      "grad_norm": 178.0490848911099,
      "kl": 0.923828125,
      "learning_rate": 1.1489187852601147e-05,
      "loss": 0.5043,
      "reward": 2.643836498260498,
      "reward_std": 0.4666139706969261,
      "rewards/accuracy_reward": 0.7447916865348816,
      "rewards/reasoning_steps_reward": 0.9756944626569748,
      "rewards/repetition_penalty_reward": -0.03237887378782034,
      "rewards/tag_count_reward": 0.9557292014360428,
      "step": 338
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 168.6822967529297,
      "epoch": 0.5084364454443194,
      "grad_norm": 33.126865780616335,
      "kl": 0.7275390625,
      "learning_rate": 1.143730513619317e-05,
      "loss": 0.2163,
      "reward": 2.5825945138931274,
      "reward_std": 0.41466938704252243,
      "rewards/accuracy_reward": 0.6666666716337204,
      "rewards/reasoning_steps_reward": 0.97743059694767,
      "rewards/repetition_penalty_reward": -0.04587779473513365,
      "rewards/tag_count_reward": 0.9843750149011612,
      "step": 339
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 199.1666717529297,
      "epoch": 0.509936257967754,
      "grad_norm": 100.89545618937568,
      "kl": 1.208984375,
      "learning_rate": 1.1385382883661881e-05,
      "loss": 0.6377,
      "reward": 2.652986526489258,
      "reward_std": 0.45409294590353966,
      "rewards/accuracy_reward": 0.7968750149011612,
      "rewards/reasoning_steps_reward": 0.958333358168602,
      "rewards/repetition_penalty_reward": -0.05144065525382757,
      "rewards/tag_count_reward": 0.9492187649011612,
      "step": 340
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 175.04687881469727,
      "epoch": 0.5114360704911886,
      "grad_norm": 294.488166826066,
      "kl": 0.87890625,
      "learning_rate": 1.1333422523238858e-05,
      "loss": 0.3826,
      "reward": 2.572603166103363,
      "reward_std": 0.42086203396320343,
      "rewards/accuracy_reward": 0.6614583432674408,
      "rewards/reasoning_steps_reward": 0.9774305820465088,
      "rewards/repetition_penalty_reward": -0.04154628235846758,
      "rewards/tag_count_reward": 0.9752604365348816,
      "step": 341
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 206.58333587646484,
      "epoch": 0.5129358830146231,
      "grad_norm": 56.33307741176176,
      "kl": 1.21875,
      "learning_rate": 1.1281425484203908e-05,
      "loss": 0.4977,
      "reward": 2.42953622341156,
      "reward_std": 0.4737429544329643,
      "rewards/accuracy_reward": 0.5572916716337204,
      "rewards/reasoning_steps_reward": 0.972222238779068,
      "rewards/repetition_penalty_reward": -0.060915243811905384,
      "rewards/tag_count_reward": 0.9609375149011612,
      "step": 342
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 166.45313262939453,
      "epoch": 0.5144356955380578,
      "grad_norm": 7.667419781427111,
      "kl": 0.6015625,
      "learning_rate": 1.122939319684577e-05,
      "loss": 0.1143,
      "reward": 2.5116847157478333,
      "reward_std": 0.43062853813171387,
      "rewards/accuracy_reward": 0.6302083507180214,
      "rewards/reasoning_steps_reward": 0.9513888955116272,
      "rewards/repetition_penalty_reward": -0.05428746622055769,
      "rewards/tag_count_reward": 0.9843750149011612,
      "step": 343
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 178.51562881469727,
      "epoch": 0.5159355080614924,
      "grad_norm": 23.5357959938352,
      "kl": 1.126953125,
      "learning_rate": 1.1177327092422761e-05,
      "loss": 0.3605,
      "reward": 2.4563730359077454,
      "reward_std": 0.4659854732453823,
      "rewards/accuracy_reward": 0.572916679084301,
      "rewards/reasoning_steps_reward": 0.9600694477558136,
      "rewards/repetition_penalty_reward": -0.04275886481627822,
      "rewards/tag_count_reward": 0.9661458432674408,
      "step": 344
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 184.5104217529297,
      "epoch": 0.5174353205849269,
      "grad_norm": 31.72517924718908,
      "kl": 0.88671875,
      "learning_rate": 1.1125228603123408e-05,
      "loss": 0.3874,
      "reward": 2.5461575388908386,
      "reward_std": 0.5080005489289761,
      "rewards/accuracy_reward": 0.6875000298023224,
      "rewards/reasoning_steps_reward": 0.9444444626569748,
      "rewards/repetition_penalty_reward": -0.05193278752267361,
      "rewards/tag_count_reward": 0.966145858168602,
      "step": 345
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 180.89062881469727,
      "epoch": 0.5189351331083615,
      "grad_norm": 87.98939762631694,
      "kl": 0.7275390625,
      "learning_rate": 1.107309916202705e-05,
      "loss": 0.381,
      "reward": 2.5558876395225525,
      "reward_std": 0.5220728367567062,
      "rewards/accuracy_reward": 0.6822916865348816,
      "rewards/reasoning_steps_reward": 0.9635416716337204,
      "rewards/repetition_penalty_reward": -0.05609158892184496,
      "rewards/tag_count_reward": 0.966145858168602,
      "step": 346
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 176.32812881469727,
      "epoch": 0.520434945631796,
      "grad_norm": 12.257830254456051,
      "kl": 0.60009765625,
      "learning_rate": 1.1020940203064425e-05,
      "loss": 0.228,
      "reward": 2.6993812322616577,
      "reward_std": 0.442532442510128,
      "rewards/accuracy_reward": 0.848958358168602,
      "rewards/reasoning_steps_reward": 0.9427083432674408,
      "rewards/repetition_penalty_reward": -0.05582709517329931,
      "rewards/tag_count_reward": 0.9635416865348816,
      "step": 347
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 204.01042938232422,
      "epoch": 0.5219347581552306,
      "grad_norm": 166.2835482202829,
      "kl": 3.4541015625,
      "learning_rate": 1.096875316097822e-05,
      "loss": 0.7129,
      "reward": 2.664002478122711,
      "reward_std": 0.5798608511686325,
      "rewards/accuracy_reward": 0.8593750149011612,
      "rewards/reasoning_steps_reward": 0.9305556118488312,
      "rewards/repetition_penalty_reward": -0.0829594787210226,
      "rewards/tag_count_reward": 0.9570312649011612,
      "step": 348
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 216.75000381469727,
      "epoch": 0.5234345706786652,
      "grad_norm": 24.78884718715676,
      "kl": 0.576171875,
      "learning_rate": 1.0916539471283607e-05,
      "loss": 0.3904,
      "reward": 2.557326078414917,
      "reward_std": 0.5754078030586243,
      "rewards/accuracy_reward": 0.723958358168602,
      "rewards/reasoning_steps_reward": 0.9687500149011612,
      "rewards/repetition_penalty_reward": -0.093715600669384,
      "rewards/tag_count_reward": 0.958333358168602,
      "step": 349
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 185.2604217529297,
      "epoch": 0.5249343832020997,
      "grad_norm": 10.924883170422843,
      "kl": 0.55810546875,
      "learning_rate": 1.0864300570228757e-05,
      "loss": 0.2519,
      "reward": 2.49711149930954,
      "reward_std": 0.5190073773264885,
      "rewards/accuracy_reward": 0.6406250223517418,
      "rewards/reasoning_steps_reward": 0.9618055522441864,
      "rewards/repetition_penalty_reward": -0.07016285322606564,
      "rewards/tag_count_reward": 0.9648437649011612,
      "step": 350
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 198.96875,
      "epoch": 0.5264341957255343,
      "grad_norm": 32.87973860027691,
      "kl": 1.43359375,
      "learning_rate": 1.0812037894755336e-05,
      "loss": 0.5601,
      "reward": 2.497571051120758,
      "reward_std": 0.710141509771347,
      "rewards/accuracy_reward": 0.692708358168602,
      "rewards/reasoning_steps_reward": 0.9236111491918564,
      "rewards/repetition_penalty_reward": -0.07577977981418371,
      "rewards/tag_count_reward": 0.9570312649011612,
      "step": 351
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 189.5885467529297,
      "epoch": 0.5279340082489689,
      "grad_norm": 35.65905039343771,
      "kl": 1.181640625,
      "learning_rate": 1.0759752882458972e-05,
      "loss": 0.4482,
      "reward": 2.4979015588760376,
      "reward_std": 0.5550749897956848,
      "rewards/accuracy_reward": 0.6562500223517418,
      "rewards/reasoning_steps_reward": 0.9565972238779068,
      "rewards/repetition_penalty_reward": -0.07067496795207262,
      "rewards/tag_count_reward": 0.9557291865348816,
      "step": 352
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 194.76563262939453,
      "epoch": 0.5294338207724034,
      "grad_norm": 37.03762904065605,
      "kl": 1.2021484375,
      "learning_rate": 1.0707446971549717e-05,
      "loss": 0.5284,
      "reward": 2.480072259902954,
      "reward_std": 0.5370368957519531,
      "rewards/accuracy_reward": 0.6510416865348816,
      "rewards/reasoning_steps_reward": 0.9409722834825516,
      "rewards/repetition_penalty_reward": -0.07938963826745749,
      "rewards/tag_count_reward": 0.9674479365348816,
      "step": 353
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 197.4010467529297,
      "epoch": 0.530933633295838,
      "grad_norm": 47.85939515897191,
      "kl": 1.6240234375,
      "learning_rate": 1.0655121600812482e-05,
      "loss": 0.3828,
      "reward": 2.545462191104889,
      "reward_std": 0.44792424887418747,
      "rewards/accuracy_reward": 0.6822916865348816,
      "rewards/reasoning_steps_reward": 0.9635416865348816,
      "rewards/repetition_penalty_reward": -0.06521502183750272,
      "rewards/tag_count_reward": 0.9648437649011612,
      "step": 354
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 197.23437881469727,
      "epoch": 0.5324334458192725,
      "grad_norm": 46.48162850362066,
      "kl": 1.5673828125,
      "learning_rate": 1.0602778209567462e-05,
      "loss": 0.501,
      "reward": 2.4815279245376587,
      "reward_std": 0.5560100227594376,
      "rewards/accuracy_reward": 0.6250000149011612,
      "rewards/reasoning_steps_reward": 0.9635416716337204,
      "rewards/repetition_penalty_reward": -0.07055553328245878,
      "rewards/tag_count_reward": 0.9635416865348816,
      "step": 355
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 195.2968864440918,
      "epoch": 0.5339332583427071,
      "grad_norm": 29.59912755896137,
      "kl": 1.427734375,
      "learning_rate": 1.0550418237630547e-05,
      "loss": 0.4327,
      "reward": 2.5716618597507477,
      "reward_std": 0.5098543167114258,
      "rewards/accuracy_reward": 0.776041679084301,
      "rewards/reasoning_steps_reward": 0.9288194924592972,
      "rewards/repetition_penalty_reward": -0.08241807296872139,
      "rewards/tag_count_reward": 0.9492187649011612,
      "step": 356
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 208.03125,
      "epoch": 0.5354330708661418,
      "grad_norm": 39.48130096076877,
      "kl": 1.7080078125,
      "learning_rate": 1.0498043125273714e-05,
      "loss": 0.6548,
      "reward": 2.4025214314460754,
      "reward_std": 0.6107548177242279,
      "rewards/accuracy_reward": 0.5781250223517418,
      "rewards/reasoning_steps_reward": 0.9375000596046448,
      "rewards/repetition_penalty_reward": -0.05190564412623644,
      "rewards/tag_count_reward": 0.9388020932674408,
      "step": 357
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 209.78125381469727,
      "epoch": 0.5369328833895763,
      "grad_norm": 38.82027339580314,
      "kl": 1.8056640625,
      "learning_rate": 1.0445654313185402e-05,
      "loss": 0.7392,
      "reward": 2.5126326084136963,
      "reward_std": 0.5534802153706551,
      "rewards/accuracy_reward": 0.6510416865348816,
      "rewards/reasoning_steps_reward": 0.9513888955116272,
      "rewards/repetition_penalty_reward": -0.04813140258193016,
      "rewards/tag_count_reward": 0.958333358168602,
      "step": 358
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 191.0729217529297,
      "epoch": 0.5384326959130109,
      "grad_norm": 24.170858033164397,
      "kl": 1.42626953125,
      "learning_rate": 1.0393253242430898e-05,
      "loss": 0.4034,
      "reward": 2.4246557354927063,
      "reward_std": 0.5046445429325104,
      "rewards/accuracy_reward": 0.5625000149011612,
      "rewards/reasoning_steps_reward": 0.9565972685813904,
      "rewards/repetition_penalty_reward": -0.04496251605451107,
      "rewards/tag_count_reward": 0.9505208432674408,
      "step": 359
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 222.46355056762695,
      "epoch": 0.5399325084364455,
      "grad_norm": 153.4349595006434,
      "kl": 2.580078125,
      "learning_rate": 1.0340841354412688e-05,
      "loss": 0.8895,
      "reward": 2.522955060005188,
      "reward_std": 0.5852703154087067,
      "rewards/accuracy_reward": 0.713541679084301,
      "rewards/reasoning_steps_reward": 0.9236111342906952,
      "rewards/repetition_penalty_reward": -0.043885353952646255,
      "rewards/tag_count_reward": 0.9296875149011612,
      "step": 360
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 234.3958396911621,
      "epoch": 0.54143232095988,
      "grad_norm": 166.56500053569175,
      "kl": 4.076171875,
      "learning_rate": 1.0288420090830803e-05,
      "loss": 1.174,
      "reward": 2.4644437432289124,
      "reward_std": 0.5811575800180435,
      "rewards/accuracy_reward": 0.6093750223517418,
      "rewards/reasoning_steps_reward": 0.9618055671453476,
      "rewards/repetition_penalty_reward": -0.04033063165843487,
      "rewards/tag_count_reward": 0.9335937649011612,
      "step": 361
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 223.4583396911621,
      "epoch": 0.5429321334833146,
      "grad_norm": 33.889969513651074,
      "kl": 1.30078125,
      "learning_rate": 1.0235990893643184e-05,
      "loss": 0.5957,
      "reward": 2.4738687872886658,
      "reward_std": 0.524218238890171,
      "rewards/accuracy_reward": 0.6093750149011612,
      "rewards/reasoning_steps_reward": 0.9618055820465088,
      "rewards/repetition_penalty_reward": -0.04913472477346659,
      "rewards/tag_count_reward": 0.9518229365348816,
      "step": 362
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 228.3645896911621,
      "epoch": 0.5444319460067492,
      "grad_norm": 60.56364684515394,
      "kl": 2.970703125,
      "learning_rate": 1.0183555205025986e-05,
      "loss": 0.6895,
      "reward": 2.5394935607910156,
      "reward_std": 0.5755364149808884,
      "rewards/accuracy_reward": 0.682291679084301,
      "rewards/reasoning_steps_reward": 0.9461805820465088,
      "rewards/repetition_penalty_reward": -0.03819744661450386,
      "rewards/tag_count_reward": 0.9492187649011612,
      "step": 363
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 202.82813262939453,
      "epoch": 0.5459317585301837,
      "grad_norm": 152.09115123663648,
      "kl": 24.283203125,
      "learning_rate": 1.0131114467333935e-05,
      "loss": 1.2772,
      "reward": 2.660370111465454,
      "reward_std": 0.5182667449116707,
      "rewards/accuracy_reward": 0.8072916865348816,
      "rewards/reasoning_steps_reward": 0.9392361044883728,
      "rewards/repetition_penalty_reward": -0.044491049367934465,
      "rewards/tag_count_reward": 0.958333358168602,
      "step": 364
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 234.3750114440918,
      "epoch": 0.5474315710536183,
      "grad_norm": 45.62424706171924,
      "kl": 6.0390625,
      "learning_rate": 1.0078670123060638e-05,
      "loss": 1.4476,
      "reward": 2.325054883956909,
      "reward_std": 0.6013178080320358,
      "rewards/accuracy_reward": 0.536458358168602,
      "rewards/reasoning_steps_reward": 0.911458358168602,
      "rewards/repetition_penalty_reward": -0.047341120429337025,
      "rewards/tag_count_reward": 0.9244791865348816,
      "step": 365
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 249.9895896911621,
      "epoch": 0.5489313835770528,
      "grad_norm": 394.34439096252885,
      "kl": 3.89453125,
      "learning_rate": 1.002622361479891e-05,
      "loss": 1.04,
      "reward": 2.2679548859596252,
      "reward_std": 0.636200875043869,
      "rewards/accuracy_reward": 0.4531250223517418,
      "rewards/reasoning_steps_reward": 0.9218750298023224,
      "rewards/repetition_penalty_reward": -0.032826476730406284,
      "rewards/tag_count_reward": 0.9257812798023224,
      "step": 366
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 228.17188262939453,
      "epoch": 0.5504311961004874,
      "grad_norm": 27.602993232186183,
      "kl": 1.998046875,
      "learning_rate": 9.973776385201093e-06,
      "loss": 0.5853,
      "reward": 2.419560134410858,
      "reward_std": 0.6460568159818649,
      "rewards/accuracy_reward": 0.5937500149011612,
      "rewards/reasoning_steps_reward": 0.9357638657093048,
      "rewards/repetition_penalty_reward": -0.04354754835367203,
      "rewards/tag_count_reward": 0.9335937649011612,
      "step": 367
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 245.15625381469727,
      "epoch": 0.551931008623922,
      "grad_norm": 31.72593423602673,
      "kl": 2.052734375,
      "learning_rate": 9.921329876939365e-06,
      "loss": 0.725,
      "reward": 2.489229917526245,
      "reward_std": 0.5430986732244492,
      "rewards/accuracy_reward": 0.6302083432674408,
      "rewards/reasoning_steps_reward": 0.9531250298023224,
      "rewards/repetition_penalty_reward": -0.04722849931567907,
      "rewards/tag_count_reward": 0.9531250149011612,
      "step": 368
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 235.01563262939453,
      "epoch": 0.5534308211473565,
      "grad_norm": 11.874407407969304,
      "kl": 1.76171875,
      "learning_rate": 9.868885532666068e-06,
      "loss": 0.6634,
      "reward": 2.2123183608055115,
      "reward_std": 0.538506917655468,
      "rewards/accuracy_reward": 0.3906250260770321,
      "rewards/reasoning_steps_reward": 0.9427083283662796,
      "rewards/repetition_penalty_reward": -0.052004692144691944,
      "rewards/tag_count_reward": 0.9309895932674408,
      "step": 369
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 211.4739646911621,
      "epoch": 0.5549306336707911,
      "grad_norm": 46.973349870839904,
      "kl": 2.095703125,
      "learning_rate": 9.816444794974018e-06,
      "loss": 0.7028,
      "reward": 2.6153268814086914,
      "reward_std": 0.5184621512889862,
      "rewards/accuracy_reward": 0.7656250298023224,
      "rewards/reasoning_steps_reward": 0.9427083432674408,
      "rewards/repetition_penalty_reward": -0.0591523889452219,
      "rewards/tag_count_reward": 0.966145858168602,
      "step": 370
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 212.7395896911621,
      "epoch": 0.5564304461942258,
      "grad_norm": 57.08569018760819,
      "kl": 16.5546875,
      "learning_rate": 9.76400910635682e-06,
      "loss": 0.7212,
      "reward": 2.2974973320961,
      "reward_std": 0.499550960958004,
      "rewards/accuracy_reward": 0.4583333432674408,
      "rewards/reasoning_steps_reward": 0.94618059694767,
      "rewards/repetition_penalty_reward": -0.045818757731467485,
      "rewards/tag_count_reward": 0.938802108168602,
      "step": 371
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 224.7239646911621,
      "epoch": 0.5579302587176603,
      "grad_norm": 18.858608822380706,
      "kl": 1.96875,
      "learning_rate": 9.7115799091692e-06,
      "loss": 0.4937,
      "reward": 2.616053879261017,
      "reward_std": 0.44517990201711655,
      "rewards/accuracy_reward": 0.7291666865348816,
      "rewards/reasoning_steps_reward": 0.9618055820465088,
      "rewards/repetition_penalty_reward": -0.03845999389886856,
      "rewards/tag_count_reward": 0.9635416865348816,
      "step": 372
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 227.83334350585938,
      "epoch": 0.5594300712410949,
      "grad_norm": 28.416379388658825,
      "kl": 3.048828125,
      "learning_rate": 9.659158645587319e-06,
      "loss": 0.9709,
      "reward": 2.659669041633606,
      "reward_std": 0.5419558137655258,
      "rewards/accuracy_reward": 0.7864583432674408,
      "rewards/reasoning_steps_reward": 0.9600694626569748,
      "rewards/repetition_penalty_reward": -0.04649423388764262,
      "rewards/tag_count_reward": 0.9596354514360428,
      "step": 373
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 235.84375381469727,
      "epoch": 0.5609298837645295,
      "grad_norm": 94.84357384526665,
      "kl": 5.3369140625,
      "learning_rate": 9.606746757569107e-06,
      "loss": 1.197,
      "reward": 2.5181403756141663,
      "reward_std": 0.6076074615120888,
      "rewards/accuracy_reward": 0.7135416716337204,
      "rewards/reasoning_steps_reward": 0.9166666716337204,
      "rewards/repetition_penalty_reward": -0.04175550863146782,
      "rewards/tag_count_reward": 0.9296875149011612,
      "step": 374
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 229.70313262939453,
      "epoch": 0.562429696287964,
      "grad_norm": 26.282174538587235,
      "kl": 1.5439453125,
      "learning_rate": 9.554345686814601e-06,
      "loss": 0.767,
      "reward": 2.4584012627601624,
      "reward_std": 0.5215486437082291,
      "rewards/accuracy_reward": 0.6093750074505806,
      "rewards/reasoning_steps_reward": 0.9548610895872116,
      "rewards/repetition_penalty_reward": -0.04854332935065031,
      "rewards/tag_count_reward": 0.942708358168602,
      "step": 375
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 221.25,
      "epoch": 0.5639295088113986,
      "grad_norm": 12.593832343409222,
      "kl": 1.1826171875,
      "learning_rate": 9.501956874726289e-06,
      "loss": 0.5623,
      "reward": 2.5942240357398987,
      "reward_std": 0.4434950575232506,
      "rewards/accuracy_reward": 0.7031250074505806,
      "rewards/reasoning_steps_reward": 0.960069477558136,
      "rewards/repetition_penalty_reward": -0.03121019806712866,
      "rewards/tag_count_reward": 0.962239608168602,
      "step": 376
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 212.0572967529297,
      "epoch": 0.5654293213348331,
      "grad_norm": 4.53299182911921,
      "kl": 1.1650390625,
      "learning_rate": 9.449581762369454e-06,
      "loss": 0.4889,
      "reward": 2.432799220085144,
      "reward_std": 0.5172925740480423,
      "rewards/accuracy_reward": 0.5416666939854622,
      "rewards/reasoning_steps_reward": 0.9670138955116272,
      "rewards/repetition_penalty_reward": -0.03942307736724615,
      "rewards/tag_count_reward": 0.9635416865348816,
      "step": 377
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 273.8177146911621,
      "epoch": 0.5669291338582677,
      "grad_norm": 148.2456214132894,
      "kl": 5.01953125,
      "learning_rate": 9.39722179043254e-06,
      "loss": 1.2941,
      "reward": 2.3988120555877686,
      "reward_std": 0.6988454312086105,
      "rewards/accuracy_reward": 0.6093750298023224,
      "rewards/reasoning_steps_reward": 0.918402835726738,
      "rewards/repetition_penalty_reward": -0.03912200313061476,
      "rewards/tag_count_reward": 0.9101562649011612,
      "step": 378
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 333.6145935058594,
      "epoch": 0.5684289463817023,
      "grad_norm": 116.2915823519784,
      "kl": 11.8125,
      "learning_rate": 9.344878399187521e-06,
      "loss": 2.0047,
      "reward": 2.2520939111709595,
      "reward_std": 0.821685403585434,
      "rewards/accuracy_reward": 0.5572916716337204,
      "rewards/reasoning_steps_reward": 0.8524305522441864,
      "rewards/repetition_penalty_reward": -0.03262835554778576,
      "rewards/tag_count_reward": 0.8750000149011612,
      "step": 379
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 420.4947967529297,
      "epoch": 0.5699287589051368,
      "grad_norm": 1197.114266206702,
      "kl": 12.4609375,
      "learning_rate": 9.292553028450286e-06,
      "loss": 1.6502,
      "reward": 2.098519653081894,
      "reward_std": 0.8776163309812546,
      "rewards/accuracy_reward": 0.5052083432674408,
      "rewards/reasoning_steps_reward": 0.8263889402151108,
      "rewards/repetition_penalty_reward": -0.0390672804787755,
      "rewards/tag_count_reward": 0.805989608168602,
      "step": 380
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 426.5677185058594,
      "epoch": 0.5714285714285714,
      "grad_norm": 26.549926170164234,
      "kl": 6.15234375,
      "learning_rate": 9.240247117541031e-06,
      "loss": 1.2117,
      "reward": 2.119348645210266,
      "reward_std": 0.8967611789703369,
      "rewards/accuracy_reward": 0.479166679084301,
      "rewards/reasoning_steps_reward": 0.84375,
      "rewards/repetition_penalty_reward": -0.02648477186448872,
      "rewards/tag_count_reward": 0.8229166865348816,
      "step": 381
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 391.78125762939453,
      "epoch": 0.572928383952006,
      "grad_norm": 88.00179700818907,
      "kl": 7.25,
      "learning_rate": 9.187962105244667e-06,
      "loss": 1.377,
      "reward": 2.291730046272278,
      "reward_std": 0.8747196942567825,
      "rewards/accuracy_reward": 0.6354166716337204,
      "rewards/reasoning_steps_reward": 0.835069477558136,
      "rewards/repetition_penalty_reward": -0.02511032810434699,
      "rewards/tag_count_reward": 0.8463542014360428,
      "step": 382
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 451.3541793823242,
      "epoch": 0.5744281964754405,
      "grad_norm": 13.361311656907796,
      "kl": 1.662109375,
      "learning_rate": 9.135699429771245e-06,
      "loss": 0.7943,
      "reward": 2.1707218885421753,
      "reward_std": 0.9110157489776611,
      "rewards/accuracy_reward": 0.5468750149011612,
      "rewards/reasoning_steps_reward": 0.8298611044883728,
      "rewards/repetition_penalty_reward": -0.02372267236933112,
      "rewards/tag_count_reward": 0.817708358168602,
      "step": 383
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 295.9948043823242,
      "epoch": 0.5759280089988752,
      "grad_norm": 30.04886642000228,
      "kl": 0.908203125,
      "learning_rate": 9.083460528716396e-06,
      "loss": 0.6667,
      "reward": 2.325647294521332,
      "reward_std": 0.6908200830221176,
      "rewards/accuracy_reward": 0.5520833395421505,
      "rewards/reasoning_steps_reward": 0.9045139253139496,
      "rewards/repetition_penalty_reward": -0.0345958243124187,
      "rewards/tag_count_reward": 0.903645858168602,
      "step": 384
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 326.80730056762695,
      "epoch": 0.5774278215223098,
      "grad_norm": 10.663780511762228,
      "kl": 0.76953125,
      "learning_rate": 9.031246839021783e-06,
      "loss": 0.6302,
      "reward": 2.3826356530189514,
      "reward_std": 0.6792797073721886,
      "rewards/accuracy_reward": 0.5937500298023224,
      "rewards/reasoning_steps_reward": 0.9340278208255768,
      "rewards/repetition_penalty_reward": -0.051392185501754284,
      "rewards/tag_count_reward": 0.9062500149011612,
      "step": 385
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 231.76041793823242,
      "epoch": 0.5789276340457443,
      "grad_norm": 61988.54428984002,
      "kl": 7898.1875,
      "learning_rate": 8.979059796935578e-06,
      "loss": 243.7071,
      "reward": 2.450250804424286,
      "reward_std": 0.6126392781734467,
      "rewards/accuracy_reward": 0.6093750149011612,
      "rewards/reasoning_steps_reward": 0.9409722238779068,
      "rewards/repetition_penalty_reward": -0.0402006134390831,
      "rewards/tag_count_reward": 0.9401041865348816,
      "step": 386
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 200.89062881469727,
      "epoch": 0.5804274465691789,
      "grad_norm": 40219.98047828671,
      "kl": 1208.828125,
      "learning_rate": 8.926900837972953e-06,
      "loss": 56.1702,
      "reward": 2.596910834312439,
      "reward_std": 0.3909384198486805,
      "rewards/accuracy_reward": 0.6822916716337204,
      "rewards/reasoning_steps_reward": 0.9774305671453476,
      "rewards/repetition_penalty_reward": -0.04197810683399439,
      "rewards/tag_count_reward": 0.9791666865348816,
      "step": 387
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 225.0260467529297,
      "epoch": 0.5819272590926134,
      "grad_norm": 21433.51255502709,
      "kl": 3312.6201171875,
      "learning_rate": 8.874771396876597e-06,
      "loss": 88.2373,
      "reward": 2.6155471205711365,
      "reward_std": 0.478760302066803,
      "rewards/accuracy_reward": 0.7343750149011612,
      "rewards/reasoning_steps_reward": 0.9687500447034836,
      "rewards/repetition_penalty_reward": -0.040702925994992256,
      "rewards/tag_count_reward": 0.9531250298023224,
      "step": 388
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 228.98958587646484,
      "epoch": 0.583427071616048,
      "grad_norm": 599372.7708149451,
      "kl": 12804.0703125,
      "learning_rate": 8.822672907577244e-06,
      "loss": 575.4508,
      "reward": 2.526801884174347,
      "reward_std": 0.5001873224973679,
      "rewards/accuracy_reward": 0.6562500149011612,
      "rewards/reasoning_steps_reward": 0.9565972536802292,
      "rewards/repetition_penalty_reward": -0.04437878727912903,
      "rewards/tag_count_reward": 0.9583333730697632,
      "step": 389
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 196.5833396911621,
      "epoch": 0.5849268841394826,
      "grad_norm": 98.60380013944835,
      "kl": 11.30810546875,
      "learning_rate": 8.770606803154235e-06,
      "loss": 0.4166,
      "reward": 2.5563814640045166,
      "reward_std": 0.3433222845196724,
      "rewards/accuracy_reward": 0.6093750223517418,
      "rewards/reasoning_steps_reward": 0.998263880610466,
      "rewards/repetition_penalty_reward": -0.04344499483704567,
      "rewards/tag_count_reward": 0.9921875149011612,
      "step": 390
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 196.7083396911621,
      "epoch": 0.5864266966629171,
      "grad_norm": 100.03328215519356,
      "kl": 8.48876953125,
      "learning_rate": 8.718574515796099e-06,
      "loss": 0.6527,
      "reward": 2.6383371353149414,
      "reward_std": 0.4097052291035652,
      "rewards/accuracy_reward": 0.723958358168602,
      "rewards/reasoning_steps_reward": 0.9791667014360428,
      "rewards/repetition_penalty_reward": -0.04395460430532694,
      "rewards/tag_count_reward": 0.9791666865348816,
      "step": 391
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 185.48437881469727,
      "epoch": 0.5879265091863517,
      "grad_norm": 4314.703902280122,
      "kl": 8.80859375,
      "learning_rate": 8.666577476761147e-06,
      "loss": 0.6093,
      "reward": 2.56997287273407,
      "reward_std": 0.39948707073926926,
      "rewards/accuracy_reward": 0.6718750298023224,
      "rewards/reasoning_steps_reward": 0.973958358168602,
      "rewards/repetition_penalty_reward": -0.04851668328046799,
      "rewards/tag_count_reward": 0.9726562649011612,
      "step": 392
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 194.4739646911621,
      "epoch": 0.5894263217097863,
      "grad_norm": 5.999735316199694,
      "kl": 0.830078125,
      "learning_rate": 8.61461711633812e-06,
      "loss": 0.2302,
      "reward": 2.6648696064949036,
      "reward_std": 0.3630646914243698,
      "rewards/accuracy_reward": 0.7552083432674408,
      "rewards/reasoning_steps_reward": 0.987847238779068,
      "rewards/repetition_penalty_reward": -0.04954020772129297,
      "rewards/tag_count_reward": 0.9713542014360428,
      "step": 393
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 187.77083587646484,
      "epoch": 0.5909261342332208,
      "grad_norm": 1.1014377641969626,
      "kl": 0.564453125,
      "learning_rate": 8.562694863806833e-06,
      "loss": 0.1169,
      "reward": 2.4198758602142334,
      "reward_std": 0.39315300434827805,
      "rewards/accuracy_reward": 0.4895833432674408,
      "rewards/reasoning_steps_reward": 0.9913194477558136,
      "rewards/repetition_penalty_reward": -0.04930820316076279,
      "rewards/tag_count_reward": 0.9882812649011612,
      "step": 394
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 222.7500114440918,
      "epoch": 0.5924259467566554,
      "grad_norm": 4.04988083724822,
      "kl": 0.7509765625,
      "learning_rate": 8.510812147398857e-06,
      "loss": 0.4591,
      "reward": 2.632589817047119,
      "reward_std": 0.5965047925710678,
      "rewards/accuracy_reward": 0.7812500149011612,
      "rewards/reasoning_steps_reward": 0.9600694626569748,
      "rewards/repetition_penalty_reward": -0.06706297304481268,
      "rewards/tag_count_reward": 0.958333358168602,
      "step": 395
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 204.34896850585938,
      "epoch": 0.59392575928009,
      "grad_norm": 2.0214381753763924,
      "kl": 0.5986328125,
      "learning_rate": 8.458970394258244e-06,
      "loss": 0.308,
      "reward": 2.5586500763893127,
      "reward_std": 0.43028920516371727,
      "rewards/accuracy_reward": 0.6562500298023224,
      "rewards/reasoning_steps_reward": 0.9774305671453476,
      "rewards/repetition_penalty_reward": -0.05289509380236268,
      "rewards/tag_count_reward": 0.9778645932674408,
      "step": 396
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 183.1979217529297,
      "epoch": 0.5954255718035245,
      "grad_norm": 0.9700944248252047,
      "kl": 0.46533203125,
      "learning_rate": 8.407171030402263e-06,
      "loss": 0.0956,
      "reward": 2.5613616704940796,
      "reward_std": 0.3627118840813637,
      "rewards/accuracy_reward": 0.6250000149011612,
      "rewards/reasoning_steps_reward": 1.0,
      "rewards/repetition_penalty_reward": -0.0545238540507853,
      "rewards/tag_count_reward": 0.9908854216337204,
      "step": 397
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 206.43750381469727,
      "epoch": 0.5969253843269592,
      "grad_norm": 3.0901812570047826,
      "kl": 0.66455078125,
      "learning_rate": 8.355415480682176e-06,
      "loss": 0.4091,
      "reward": 2.456976294517517,
      "reward_std": 0.3992188200354576,
      "rewards/accuracy_reward": 0.5520833432674408,
      "rewards/reasoning_steps_reward": 0.9826389104127884,
      "rewards/repetition_penalty_reward": -0.05300639010965824,
      "rewards/tag_count_reward": 0.9752604365348816,
      "step": 398
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 205.4322967529297,
      "epoch": 0.5984251968503937,
      "grad_norm": 0.9560339346770327,
      "kl": 0.677734375,
      "learning_rate": 8.303705168744042e-06,
      "loss": 0.386,
      "reward": 2.42321240901947,
      "reward_std": 0.4326848238706589,
      "rewards/accuracy_reward": 0.5052083432674408,
      "rewards/reasoning_steps_reward": 0.9774305820465088,
      "rewards/repetition_penalty_reward": -0.03338485397398472,
      "rewards/tag_count_reward": 0.9739583432674408,
      "step": 399
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 223.17708587646484,
      "epoch": 0.5999250093738283,
      "grad_norm": 3.695423652449584,
      "kl": 1.115234375,
      "learning_rate": 8.252041516989565e-06,
      "loss": 0.7209,
      "reward": 2.4974151253700256,
      "reward_std": 0.5143779292702675,
      "rewards/accuracy_reward": 0.5937500223517418,
      "rewards/reasoning_steps_reward": 0.9704861342906952,
      "rewards/repetition_penalty_reward": -0.02775857038795948,
      "rewards/tag_count_reward": 0.9609375298023224,
      "step": 400
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 278.4583435058594,
      "epoch": 0.6014248218972629,
      "grad_norm": 1.4515991916390434,
      "kl": 1.1875,
      "learning_rate": 8.200425946536956e-06,
      "loss": 0.9453,
      "reward": 2.54727965593338,
      "reward_std": 0.6352264881134033,
      "rewards/accuracy_reward": 0.7083333432674408,
      "rewards/reasoning_steps_reward": 0.9409722685813904,
      "rewards/repetition_penalty_reward": -0.023900966625660658,
      "rewards/tag_count_reward": 0.9218750149011612,
      "step": 401
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 253.0260467529297,
      "epoch": 0.6029246344206974,
      "grad_norm": 1.6992040499230678,
      "kl": 1.1376953125,
      "learning_rate": 8.148859877181849e-06,
      "loss": 0.6776,
      "reward": 2.558239758014679,
      "reward_std": 0.5012032613158226,
      "rewards/accuracy_reward": 0.6927083432674408,
      "rewards/reasoning_steps_reward": 0.951388880610466,
      "rewards/repetition_penalty_reward": -0.02205547597259283,
      "rewards/tag_count_reward": 0.9361979514360428,
      "step": 402
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 280.1458435058594,
      "epoch": 0.604424446944132,
      "grad_norm": 1.2357512902908514,
      "kl": 1.0625,
      "learning_rate": 8.097344727358247e-06,
      "loss": 0.6856,
      "reward": 2.4696335196495056,
      "reward_std": 0.6513710990548134,
      "rewards/accuracy_reward": 0.6197916716337204,
      "rewards/reasoning_steps_reward": 0.9444444626569748,
      "rewards/repetition_penalty_reward": -0.019081805367022753,
      "rewards/tag_count_reward": 0.9244791865348816,
      "step": 403
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 218.3802146911621,
      "epoch": 0.6059242594675666,
      "grad_norm": 2.5806348482702632,
      "kl": 0.6376953125,
      "learning_rate": 8.045881914099503e-06,
      "loss": 0.6086,
      "reward": 2.3685402274131775,
      "reward_std": 0.4312494173645973,
      "rewards/accuracy_reward": 0.463541679084301,
      "rewards/reasoning_steps_reward": 0.967013880610466,
      "rewards/repetition_penalty_reward": -0.024254921358078718,
      "rewards/tag_count_reward": 0.962239608168602,
      "step": 404
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 214.18750381469727,
      "epoch": 0.6074240719910011,
      "grad_norm": 3.0889995708717803,
      "kl": 0.802734375,
      "learning_rate": 7.99447285299934e-06,
      "loss": 0.9386,
      "reward": 2.6483540534973145,
      "reward_std": 0.57868642359972,
      "rewards/accuracy_reward": 0.7760416865348816,
      "rewards/reasoning_steps_reward": 0.9357639253139496,
      "rewards/repetition_penalty_reward": -0.012670394266024232,
      "rewards/tag_count_reward": 0.9492187798023224,
      "step": 405
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 255.90625381469727,
      "epoch": 0.6089238845144357,
      "grad_norm": 6.886520330181463,
      "kl": 1.41796875,
      "learning_rate": 7.943118958172917e-06,
      "loss": 0.9431,
      "reward": 2.4443989396095276,
      "reward_std": 0.7025687843561172,
      "rewards/accuracy_reward": 0.6562500149011612,
      "rewards/reasoning_steps_reward": 0.8854166865348816,
      "rewards/repetition_penalty_reward": -0.01914279453922063,
      "rewards/tag_count_reward": 0.9218750149011612,
      "step": 406
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 270.8437614440918,
      "epoch": 0.6104236970378702,
      "grad_norm": 4.059148084337778,
      "kl": 1.1845703125,
      "learning_rate": 7.891821642217926e-06,
      "loss": 0.9331,
      "reward": 2.216310352087021,
      "reward_std": 0.7253094911575317,
      "rewards/accuracy_reward": 0.442708358168602,
      "rewards/reasoning_steps_reward": 0.8854167014360428,
      "rewards/repetition_penalty_reward": -0.02066887845285237,
      "rewards/tag_count_reward": 0.9088541865348816,
      "step": 407
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 251.8072967529297,
      "epoch": 0.6119235095613048,
      "grad_norm": 3.615908850384094,
      "kl": 0.9541015625,
      "learning_rate": 7.840582316175737e-06,
      "loss": 0.7562,
      "reward": 2.1759954690933228,
      "reward_std": 0.7425311505794525,
      "rewards/accuracy_reward": 0.3958333432674408,
      "rewards/reasoning_steps_reward": 0.892361119389534,
      "rewards/repetition_penalty_reward": -0.01975110382772982,
      "rewards/tag_count_reward": 0.907552108168602,
      "step": 408
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 223.29687881469727,
      "epoch": 0.6134233220847394,
      "grad_norm": 7.82629047907828,
      "kl": 0.95703125,
      "learning_rate": 7.789402389492582e-06,
      "loss": 1.0177,
      "reward": 2.383803129196167,
      "reward_std": 0.6715415120124817,
      "rewards/accuracy_reward": 0.5520833507180214,
      "rewards/reasoning_steps_reward": 0.9184027910232544,
      "rewards/repetition_penalty_reward": -0.013766373042017221,
      "rewards/tag_count_reward": 0.9270833432674408,
      "step": 409
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 244.07292556762695,
      "epoch": 0.6149231346081739,
      "grad_norm": 7.0703944532697145,
      "kl": 0.80419921875,
      "learning_rate": 7.738283269980798e-06,
      "loss": 0.8963,
      "reward": 2.4084761142730713,
      "reward_std": 0.7272981628775597,
      "rewards/accuracy_reward": 0.6197916716337204,
      "rewards/reasoning_steps_reward": 0.8958333730697632,
      "rewards/repetition_penalty_reward": -0.018607289995998144,
      "rewards/tag_count_reward": 0.9114583432674408,
      "step": 410
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 174.3489646911621,
      "epoch": 0.6164229471316085,
      "grad_norm": 3.740634216779511,
      "kl": 0.63427734375,
      "learning_rate": 7.687226363780084e-06,
      "loss": 0.5151,
      "reward": 2.5853647589683533,
      "reward_std": 0.46327926218509674,
      "rewards/accuracy_reward": 0.677083358168602,
      "rewards/reasoning_steps_reward": 0.9618055820465088,
      "rewards/repetition_penalty_reward": -0.023576234467327595,
      "rewards/tag_count_reward": 0.970052108168602,
      "step": 411
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 205.3177146911621,
      "epoch": 0.6179227596550432,
      "grad_norm": 2.7293051628640344,
      "kl": 0.50927734375,
      "learning_rate": 7.636233075318824e-06,
      "loss": 0.6524,
      "reward": 2.4624382853507996,
      "reward_std": 0.5226383320987225,
      "rewards/accuracy_reward": 0.5989583432674408,
      "rewards/reasoning_steps_reward": 0.9340278059244156,
      "rewards/repetition_penalty_reward": -0.017162481555715203,
      "rewards/tag_count_reward": 0.9466145932674408,
      "step": 412
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 208.1822967529297,
      "epoch": 0.6194225721784777,
      "grad_norm": 14.252860190179556,
      "kl": 0.8994140625,
      "learning_rate": 7.585304807275473e-06,
      "loss": 0.7206,
      "reward": 2.4753499627113342,
      "reward_std": 0.536663368344307,
      "rewards/accuracy_reward": 0.5989583507180214,
      "rewards/reasoning_steps_reward": 0.9409722238779068,
      "rewards/repetition_penalty_reward": -0.01510148635134101,
      "rewards/tag_count_reward": 0.950520858168602,
      "step": 413
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 174.04687881469727,
      "epoch": 0.6209223847019123,
      "grad_norm": 2.8964016768843623,
      "kl": 0.73388671875,
      "learning_rate": 7.534442960539956e-06,
      "loss": 0.4618,
      "reward": 2.6283841133117676,
      "reward_std": 0.4127518758177757,
      "rewards/accuracy_reward": 0.7031250223517418,
      "rewards/reasoning_steps_reward": 0.9670139253139496,
      "rewards/repetition_penalty_reward": -0.011807008180767298,
      "rewards/tag_count_reward": 0.970052108168602,
      "step": 414
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 184.96875762939453,
      "epoch": 0.6224221972253469,
      "grad_norm": 5.401451233630221,
      "kl": 0.54443359375,
      "learning_rate": 7.483648934175138e-06,
      "loss": 0.5382,
      "reward": 2.485370635986328,
      "reward_std": 0.4186149761080742,
      "rewards/accuracy_reward": 0.5833333432674408,
      "rewards/reasoning_steps_reward": 0.9600694626569748,
      "rewards/repetition_penalty_reward": -0.020271844463422894,
      "rewards/tag_count_reward": 0.962239608168602,
      "step": 415
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 156.63021087646484,
      "epoch": 0.6239220097487814,
      "grad_norm": 12.179499233084556,
      "kl": 0.55126953125,
      "learning_rate": 7.432924125378345e-06,
      "loss": 0.4525,
      "reward": 2.612027883529663,
      "reward_std": 0.34135545045137405,
      "rewards/accuracy_reward": 0.6875000149011612,
      "rewards/reasoning_steps_reward": 0.9687500298023224,
      "rewards/repetition_penalty_reward": -0.011670089792460203,
      "rewards/tag_count_reward": 0.9674479514360428,
      "step": 416
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 185.3802146911621,
      "epoch": 0.625421822272216,
      "grad_norm": 27.64718995248203,
      "kl": 0.61962890625,
      "learning_rate": 7.382269929442925e-06,
      "loss": 0.3543,
      "reward": 2.6146509051322937,
      "reward_std": 0.3778613116592169,
      "rewards/accuracy_reward": 0.6979166716337204,
      "rewards/reasoning_steps_reward": 0.9739583432674408,
      "rewards/repetition_penalty_reward": -0.01425547618418932,
      "rewards/tag_count_reward": 0.9570312649011612,
      "step": 417
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 200.8333396911621,
      "epoch": 0.6269216347956506,
      "grad_norm": 35.112382210723155,
      "kl": 0.74609375,
      "learning_rate": 7.331687739719868e-06,
      "loss": 0.3159,
      "reward": 2.7482373118400574,
      "reward_std": 0.4861888214945793,
      "rewards/accuracy_reward": 0.8750000149011612,
      "rewards/reasoning_steps_reward": 0.9618055820465088,
      "rewards/repetition_penalty_reward": -0.03388085588812828,
      "rewards/tag_count_reward": 0.9453125298023224,
      "step": 418
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 186.5416717529297,
      "epoch": 0.6284214473190851,
      "grad_norm": 9.808027472374857,
      "kl": 0.75390625,
      "learning_rate": 7.281178947579484e-06,
      "loss": 0.3908,
      "reward": 2.5124480724334717,
      "reward_std": 0.39367250353097916,
      "rewards/accuracy_reward": 0.6145833432674408,
      "rewards/reasoning_steps_reward": 0.9895833432674408,
      "rewards/repetition_penalty_reward": -0.050051978789269924,
      "rewards/tag_count_reward": 0.958333358168602,
      "step": 419
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 210.65625762939453,
      "epoch": 0.6299212598425197,
      "grad_norm": 316.11143834667564,
      "kl": 6.1318359375,
      "learning_rate": 7.230744942373125e-06,
      "loss": 1.196,
      "reward": 2.612563908100128,
      "reward_std": 0.5070172511041164,
      "rewards/accuracy_reward": 0.7447916865348816,
      "rewards/reasoning_steps_reward": 0.9791666865348816,
      "rewards/repetition_penalty_reward": -0.06451944634318352,
      "rewards/tag_count_reward": 0.953125,
      "step": 420
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 167.5520896911621,
      "epoch": 0.6314210723659542,
      "grad_norm": 2256.612496535129,
      "kl": 120.529296875,
      "learning_rate": 7.1803871113949675e-06,
      "loss": 5.9692,
      "reward": 2.5741260647773743,
      "reward_std": 0.364616222679615,
      "rewards/accuracy_reward": 0.645833358168602,
      "rewards/reasoning_steps_reward": 0.9843750149011612,
      "rewards/repetition_penalty_reward": -0.03134271129965782,
      "rewards/tag_count_reward": 0.9752604365348816,
      "step": 421
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 210.5885467529297,
      "epoch": 0.6329208848893888,
      "grad_norm": 31650.105311482544,
      "kl": 379.5390625,
      "learning_rate": 7.13010683984386e-06,
      "loss": 45.4656,
      "reward": 2.404063105583191,
      "reward_std": 0.6292674243450165,
      "rewards/accuracy_reward": 0.5833333507180214,
      "rewards/reasoning_steps_reward": 0.9565972536802292,
      "rewards/repetition_penalty_reward": -0.07336751371622086,
      "rewards/tag_count_reward": 0.9375000149011612,
      "step": 422
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 195.60937881469727,
      "epoch": 0.6344206974128234,
      "grad_norm": 69.45154521104752,
      "kl": 2.13525390625,
      "learning_rate": 7.07990551078521e-06,
      "loss": 0.6039,
      "reward": 2.7996047139167786,
      "reward_std": 0.37217236310243607,
      "rewards/accuracy_reward": 0.8854166716337204,
      "rewards/reasoning_steps_reward": 0.9843750149011612,
      "rewards/repetition_penalty_reward": -0.03633271250873804,
      "rewards/tag_count_reward": 0.9661458432674408,
      "step": 423
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 221.22917938232422,
      "epoch": 0.6359205099362579,
      "grad_norm": 63.609726332225755,
      "kl": 2.5576171875,
      "learning_rate": 7.029784505112948e-06,
      "loss": 0.7597,
      "reward": 2.4892146587371826,
      "reward_std": 0.55179613083601,
      "rewards/accuracy_reward": 0.6406250149011612,
      "rewards/reasoning_steps_reward": 0.9618055820465088,
      "rewards/repetition_penalty_reward": -0.05332011543214321,
      "rewards/tag_count_reward": 0.9401042014360428,
      "step": 424
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 196.7864646911621,
      "epoch": 0.6374203224596925,
      "grad_norm": 30.322718341900693,
      "kl": 0.7666015625,
      "learning_rate": 6.979745201511531e-06,
      "loss": 0.5128,
      "reward": 2.699239432811737,
      "reward_std": 0.4398190379142761,
      "rewards/accuracy_reward": 0.8020833432674408,
      "rewards/reasoning_steps_reward": 0.973958358168602,
      "rewards/repetition_penalty_reward": -0.031229355605319142,
      "rewards/tag_count_reward": 0.954427108168602,
      "step": 425
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 175.7291717529297,
      "epoch": 0.6389201349831272,
      "grad_norm": 763.6392481614376,
      "kl": 442.4111328125,
      "learning_rate": 6.929788976418044e-06,
      "loss": 1.9958,
      "reward": 2.833931624889374,
      "reward_std": 0.3742608136963099,
      "rewards/accuracy_reward": 0.9062500149011612,
      "rewards/reasoning_steps_reward": 0.973958358168602,
      "rewards/repetition_penalty_reward": -0.0215371900703758,
      "rewards/tag_count_reward": 0.9752604365348816,
      "step": 426
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 226.6927146911621,
      "epoch": 0.6404199475065617,
      "grad_norm": 11.61389336470008,
      "kl": 0.611328125,
      "learning_rate": 6.879917203984306e-06,
      "loss": 0.4047,
      "reward": 2.439822733402252,
      "reward_std": 0.40970153361558914,
      "rewards/accuracy_reward": 0.5520833507180214,
      "rewards/reasoning_steps_reward": 0.984375,
      "rewards/repetition_penalty_reward": -0.05366701539605856,
      "rewards/tag_count_reward": 0.9570312798023224,
      "step": 427
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 212.5989646911621,
      "epoch": 0.6419197600299963,
      "grad_norm": 8.516759254331243,
      "kl": 0.5595703125,
      "learning_rate": 6.830131256039094e-06,
      "loss": 0.4026,
      "reward": 2.511458396911621,
      "reward_std": 0.4569522365927696,
      "rewards/accuracy_reward": 0.614583358168602,
      "rewards/reasoning_steps_reward": 0.9895833432674408,
      "rewards/repetition_penalty_reward": -0.058854201808571815,
      "rewards/tag_count_reward": 0.966145858168602,
      "step": 428
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 193.1458396911621,
      "epoch": 0.6434195725534309,
      "grad_norm": 3.0244823017102758,
      "kl": 0.5078125,
      "learning_rate": 6.7804325020504e-06,
      "loss": 0.3328,
      "reward": 2.6280587315559387,
      "reward_std": 0.33434533327817917,
      "rewards/accuracy_reward": 0.6875000298023224,
      "rewards/reasoning_steps_reward": 0.9947916716337204,
      "rewards/repetition_penalty_reward": -0.04251436982303858,
      "rewards/tag_count_reward": 0.9882812798023224,
      "step": 429
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 213.2708396911621,
      "epoch": 0.6449193850768654,
      "grad_norm": 54.16361365177273,
      "kl": 1.1318359375,
      "learning_rate": 6.730822309087756e-06,
      "loss": 0.5304,
      "reward": 2.617822825908661,
      "reward_std": 0.5841463133692741,
      "rewards/accuracy_reward": 0.7343750298023224,
      "rewards/reasoning_steps_reward": 0.9739583432674408,
      "rewards/repetition_penalty_reward": -0.061864775605499744,
      "rewards/tag_count_reward": 0.9713541865348816,
      "step": 430
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 233.47396087646484,
      "epoch": 0.6464191976003,
      "grad_norm": 9.396207453842536,
      "kl": 0.5869140625,
      "learning_rate": 6.6813020417846456e-06,
      "loss": 0.5755,
      "reward": 2.444421410560608,
      "reward_std": 0.732051394879818,
      "rewards/accuracy_reward": 0.630208358168602,
      "rewards/reasoning_steps_reward": 0.9479166567325592,
      "rewards/repetition_penalty_reward": -0.08552652504295111,
      "rewards/tag_count_reward": 0.9518229514360428,
      "step": 431
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 217.7604217529297,
      "epoch": 0.6479190101237345,
      "grad_norm": 104.85884344998551,
      "kl": 0.8720703125,
      "learning_rate": 6.6318730623009465e-06,
      "loss": 0.7651,
      "reward": 2.6597663164138794,
      "reward_std": 0.7930338382720947,
      "rewards/accuracy_reward": 0.8333333432674408,
      "rewards/reasoning_steps_reward": 0.9496527910232544,
      "rewards/repetition_penalty_reward": -0.07634495198726654,
      "rewards/tag_count_reward": 0.9531250149011612,
      "step": 432
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 194.21875381469727,
      "epoch": 0.6494188226471691,
      "grad_norm": 4283739.768658449,
      "kl": 21367.5,
      "learning_rate": 6.582536730285476e-06,
      "loss": 4439.4258,
      "reward": 2.63147896528244,
      "reward_std": 0.7544302493333817,
      "rewards/accuracy_reward": 0.833333358168602,
      "rewards/reasoning_steps_reward": 0.928819477558136,
      "rewards/repetition_penalty_reward": -0.07338221184909344,
      "rewards/tag_count_reward": 0.9427083730697632,
      "step": 433
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 235.91667556762695,
      "epoch": 0.6509186351706037,
      "grad_norm": 210786.70991777242,
      "kl": 8894.5,
      "learning_rate": 6.5332944028385885e-06,
      "loss": 447.474,
      "reward": 2.3411070704460144,
      "reward_std": 0.8666418790817261,
      "rewards/accuracy_reward": 0.6093750149011612,
      "rewards/reasoning_steps_reward": 0.9149305820465088,
      "rewards/repetition_penalty_reward": -0.10637565143406391,
      "rewards/tag_count_reward": 0.923177108168602,
      "step": 434
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 223.79687881469727,
      "epoch": 0.6524184476940382,
      "grad_norm": 1518.3629252174667,
      "kl": 64.171875,
      "learning_rate": 6.484147434474837e-06,
      "loss": 5.1614,
      "reward": 2.2417006492614746,
      "reward_std": 0.8650868535041809,
      "rewards/accuracy_reward": 0.5104166865348816,
      "rewards/reasoning_steps_reward": 0.9114583432674408,
      "rewards/repetition_penalty_reward": -0.09814327582716942,
      "rewards/tag_count_reward": 0.9179687649011612,
      "step": 435
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 220.6197967529297,
      "epoch": 0.6539182602174728,
      "grad_norm": 50.76737216477434,
      "kl": 0.66455078125,
      "learning_rate": 6.435097177085728e-06,
      "loss": 0.6146,
      "reward": 2.2728312611579895,
      "reward_std": 0.8381006345152855,
      "rewards/accuracy_reward": 0.5052083432674408,
      "rewards/reasoning_steps_reward": 0.9218750298023224,
      "rewards/repetition_penalty_reward": -0.08914803247898817,
      "rewards/tag_count_reward": 0.934895858168602,
      "step": 436
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 190.6458396911621,
      "epoch": 0.6554180727409074,
      "grad_norm": 27.797869724896955,
      "kl": 0.7705078125,
      "learning_rate": 6.386144979902527e-06,
      "loss": 0.3753,
      "reward": 2.3050162196159363,
      "reward_std": 0.77731654047966,
      "rewards/accuracy_reward": 0.5208333432674408,
      "rewards/reasoning_steps_reward": 0.927083358168602,
      "rewards/repetition_penalty_reward": -0.07909846305847168,
      "rewards/tag_count_reward": 0.9361979365348816,
      "step": 437
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 230.3697967529297,
      "epoch": 0.6569178852643419,
      "grad_norm": 117.99768545242851,
      "kl": 1.0126953125,
      "learning_rate": 6.337292189459139e-06,
      "loss": 0.6907,
      "reward": 2.48311048746109,
      "reward_std": 0.833274632692337,
      "rewards/accuracy_reward": 0.7447916865348816,
      "rewards/reasoning_steps_reward": 0.91493059694767,
      "rewards/repetition_penalty_reward": -0.10109082609415054,
      "rewards/tag_count_reward": 0.9244791716337204,
      "step": 438
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 208.05208587646484,
      "epoch": 0.6584176977877765,
      "grad_norm": 2775.548898191414,
      "kl": 79.640625,
      "learning_rate": 6.2885401495550826e-06,
      "loss": 10.0176,
      "reward": 2.4023959636688232,
      "reward_std": 0.7233280688524246,
      "rewards/accuracy_reward": 0.5937500298023224,
      "rewards/reasoning_steps_reward": 0.935763880610466,
      "rewards/repetition_penalty_reward": -0.0737327765673399,
      "rewards/tag_count_reward": 0.9466145932674408,
      "step": 439
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 211.5260467529297,
      "epoch": 0.6599175103112112,
      "grad_norm": 19900.674181692015,
      "kl": 446.0,
      "learning_rate": 6.239890201218517e-06,
      "loss": 65.0693,
      "reward": 2.270346522331238,
      "reward_std": 0.8372077494859695,
      "rewards/accuracy_reward": 0.5208333507180214,
      "rewards/reasoning_steps_reward": 0.9114583730697632,
      "rewards/repetition_penalty_reward": -0.08512232266366482,
      "rewards/tag_count_reward": 0.923177108168602,
      "step": 440
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 213.95313262939453,
      "epoch": 0.6614173228346457,
      "grad_norm": 1172.7691458595361,
      "kl": 36.59375,
      "learning_rate": 6.191343682669357e-06,
      "loss": 5.7344,
      "reward": 2.4517070651054382,
      "reward_std": 0.7322164475917816,
      "rewards/accuracy_reward": 0.6406250074505806,
      "rewards/reasoning_steps_reward": 0.944444477558136,
      "rewards/repetition_penalty_reward": -0.07867502607405186,
      "rewards/tag_count_reward": 0.9453125149011612,
      "step": 441
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 250.46355056762695,
      "epoch": 0.6629171353580803,
      "grad_norm": 105.35481274422847,
      "kl": 1.66796875,
      "learning_rate": 6.142901929282459e-06,
      "loss": 0.852,
      "reward": 2.3720561265945435,
      "reward_std": 0.936857059597969,
      "rewards/accuracy_reward": 0.6354166865348816,
      "rewards/reasoning_steps_reward": 0.9166666865348816,
      "rewards/repetition_penalty_reward": -0.10190227814018726,
      "rewards/tag_count_reward": 0.921875,
      "step": 442
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 229.20313262939453,
      "epoch": 0.6644169478815148,
      "grad_norm": 95.75169686366347,
      "kl": 1.2333984375,
      "learning_rate": 6.094566273550899e-06,
      "loss": 0.6136,
      "reward": 2.508533537387848,
      "reward_std": 0.9894896894693375,
      "rewards/accuracy_reward": 0.8072917014360428,
      "rewards/reasoning_steps_reward": 0.8958333432674408,
      "rewards/repetition_penalty_reward": -0.09823731146752834,
      "rewards/tag_count_reward": 0.9036458432674408,
      "step": 443
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 287.8229217529297,
      "epoch": 0.6659167604049494,
      "grad_norm": 62.198954440927515,
      "kl": 3.91796875,
      "learning_rate": 6.046338045049307e-06,
      "loss": 1.0033,
      "reward": 2.1896302700042725,
      "reward_std": 1.0947272032499313,
      "rewards/accuracy_reward": 0.5677083730697632,
      "rewards/reasoning_steps_reward": 0.8802083432674408,
      "rewards/repetition_penalty_reward": -0.14760945178568363,
      "rewards/tag_count_reward": 0.8893229216337204,
      "step": 444
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 210.73438262939453,
      "epoch": 0.667416572928384,
      "grad_norm": 624.3572181079038,
      "kl": 32.46875,
      "learning_rate": 5.998218570397298e-06,
      "loss": 3.7004,
      "reward": 2.489894211292267,
      "reward_std": 0.7204201519489288,
      "rewards/accuracy_reward": 0.6979166865348816,
      "rewards/reasoning_steps_reward": 0.9340277761220932,
      "rewards/repetition_penalty_reward": -0.08345651999115944,
      "rewards/tag_count_reward": 0.9414062649011612,
      "step": 445
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 192.6927146911621,
      "epoch": 0.6689163854518185,
      "grad_norm": 1291.0531241853187,
      "kl": 58.40625,
      "learning_rate": 5.950209173222985e-06,
      "loss": 6.5055,
      "reward": 2.347103238105774,
      "reward_std": 1.0178454369306564,
      "rewards/accuracy_reward": 0.666666679084301,
      "rewards/reasoning_steps_reward": 0.8715278059244156,
      "rewards/repetition_penalty_reward": -0.08432047069072723,
      "rewards/tag_count_reward": 0.8932291865348816,
      "step": 446
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 188.81250381469727,
      "epoch": 0.6704161979752531,
      "grad_norm": 167.1713247377813,
      "kl": 12.703125,
      "learning_rate": 5.902311174126565e-06,
      "loss": 1.3266,
      "reward": 2.487611711025238,
      "reward_std": 0.847619041800499,
      "rewards/accuracy_reward": 0.723958358168602,
      "rewards/reasoning_steps_reward": 0.9184027910232544,
      "rewards/repetition_penalty_reward": -0.07792646810412407,
      "rewards/tag_count_reward": 0.9231770932674408,
      "step": 447
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 217.70312881469727,
      "epoch": 0.6719160104986877,
      "grad_norm": 46.56836347829574,
      "kl": 2.587890625,
      "learning_rate": 5.854525890643996e-06,
      "loss": 0.5503,
      "reward": 2.271707057952881,
      "reward_std": 0.9676771610975266,
      "rewards/accuracy_reward": 0.583333358168602,
      "rewards/reasoning_steps_reward": 0.880208358168602,
      "rewards/repetition_penalty_reward": -0.09678260423243046,
      "rewards/tag_count_reward": 0.9049479365348816,
      "step": 448
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 212.70833587646484,
      "epoch": 0.6734158230221222,
      "grad_norm": 58.535345861262144,
      "kl": 1.36328125,
      "learning_rate": 5.806854637210752e-06,
      "loss": 0.4487,
      "reward": 2.4692699909210205,
      "reward_std": 1.0627684146165848,
      "rewards/accuracy_reward": 0.7812500149011612,
      "rewards/reasoning_steps_reward": 0.8802083730697632,
      "rewards/repetition_penalty_reward": -0.09583419561386108,
      "rewards/tag_count_reward": 0.9036458432674408,
      "step": 449
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 217.21875381469727,
      "epoch": 0.6749156355455568,
      "grad_norm": 44.853870574601515,
      "kl": 2.22265625,
      "learning_rate": 5.759298725125671e-06,
      "loss": 0.57,
      "reward": 2.438233256340027,
      "reward_std": 0.9969596564769745,
      "rewards/accuracy_reward": 0.739583358168602,
      "rewards/reasoning_steps_reward": 0.8975694626569748,
      "rewards/repetition_penalty_reward": -0.09475286118686199,
      "rewards/tag_count_reward": 0.895833358168602,
      "step": 450
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 181.16146850585938,
      "epoch": 0.6764154480689913,
      "grad_norm": 45.94535859885079,
      "kl": 5.15234375,
      "learning_rate": 5.711859462514883e-06,
      "loss": 0.6531,
      "reward": 2.3718193769454956,
      "reward_std": 0.8176105469465256,
      "rewards/accuracy_reward": 0.5937500149011612,
      "rewards/reasoning_steps_reward": 0.9149305671453476,
      "rewards/repetition_penalty_reward": -0.05483006127178669,
      "rewards/tag_count_reward": 0.9179687649011612,
      "step": 451
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 183.48958587646484,
      "epoch": 0.6779152605924259,
      "grad_norm": 54.34402687729516,
      "kl": 7.81640625,
      "learning_rate": 5.664538154295827e-06,
      "loss": 0.6031,
      "reward": 2.3453221917152405,
      "reward_std": 0.8092229068279266,
      "rewards/accuracy_reward": 0.604166679084301,
      "rewards/reasoning_steps_reward": 0.8906250149011612,
      "rewards/repetition_penalty_reward": -0.05702169891446829,
      "rewards/tag_count_reward": 0.907552108168602,
      "step": 452
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 177.88541793823242,
      "epoch": 0.6794150731158605,
      "grad_norm": 8.683387363521502,
      "kl": 2.787109375,
      "learning_rate": 5.617336102141356e-06,
      "loss": 0.3205,
      "reward": 2.4910064935684204,
      "reward_std": 0.6290017366409302,
      "rewards/accuracy_reward": 0.6354166865348816,
      "rewards/reasoning_steps_reward": 0.9531250149011612,
      "rewards/repetition_penalty_reward": -0.04545190371572971,
      "rewards/tag_count_reward": 0.9479166865348816,
      "step": 453
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 178.65625381469727,
      "epoch": 0.6809148856392951,
      "grad_norm": 21.261847254596837,
      "kl": 1.23828125,
      "learning_rate": 5.570254604443929e-06,
      "loss": 0.0513,
      "reward": 2.4800949692726135,
      "reward_std": 0.9092780351638794,
      "rewards/accuracy_reward": 0.7083333432674408,
      "rewards/reasoning_steps_reward": 0.9149305820465088,
      "rewards/repetition_penalty_reward": -0.052023096941411495,
      "rewards/tag_count_reward": 0.9088541865348816,
      "step": 454
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 189.70833587646484,
      "epoch": 0.6824146981627297,
      "grad_norm": 32.86231011635845,
      "kl": 0.9814453125,
      "learning_rate": 5.5232949562799055e-06,
      "loss": 0.2151,
      "reward": 2.370435267686844,
      "reward_std": 0.8025718629360199,
      "rewards/accuracy_reward": 0.6041666939854622,
      "rewards/reasoning_steps_reward": 0.911458358168602,
      "rewards/repetition_penalty_reward": -0.05143989436328411,
      "rewards/tag_count_reward": 0.9062500298023224,
      "step": 455
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 191.734375,
      "epoch": 0.6839145106861643,
      "grad_norm": 17.49572401123107,
      "kl": 1.21484375,
      "learning_rate": 5.4764584493739095e-06,
      "loss": 0.0338,
      "reward": 2.206531524658203,
      "reward_std": 0.7957804501056671,
      "rewards/accuracy_reward": 0.4375000223517418,
      "rewards/reasoning_steps_reward": 0.913194477558136,
      "rewards/repetition_penalty_reward": -0.05171508714556694,
      "rewards/tag_count_reward": 0.907552108168602,
      "step": 456
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 185.22916793823242,
      "epoch": 0.6854143232095988,
      "grad_norm": 12.624756226573414,
      "kl": 2.28515625,
      "learning_rate": 5.429746372063309e-06,
      "loss": 0.2271,
      "reward": 2.3943753242492676,
      "reward_std": 0.7135017514228821,
      "rewards/accuracy_reward": 0.5885416716337204,
      "rewards/reasoning_steps_reward": 0.9270833432674408,
      "rewards/repetition_penalty_reward": -0.04703101795166731,
      "rewards/tag_count_reward": 0.9257812798023224,
      "step": 457
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 193.81250381469727,
      "epoch": 0.6869141357330334,
      "grad_norm": 105.8474864914367,
      "kl": 8.5908203125,
      "learning_rate": 5.3831600092627704e-06,
      "loss": 1.1342,
      "reward": 2.7418383955955505,
      "reward_std": 0.4850631505250931,
      "rewards/accuracy_reward": 0.864583358168602,
      "rewards/reasoning_steps_reward": 0.9670139104127884,
      "rewards/repetition_penalty_reward": -0.04418608546257019,
      "rewards/tag_count_reward": 0.954427108168602,
      "step": 458
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 200.70833587646484,
      "epoch": 0.688413948256468,
      "grad_norm": 87.25261858275486,
      "kl": 10.28515625,
      "learning_rate": 5.336700642428913e-06,
      "loss": 0.6609,
      "reward": 2.498172700405121,
      "reward_std": 0.5890463814139366,
      "rewards/accuracy_reward": 0.5937500223517418,
      "rewards/reasoning_steps_reward": 0.973958358168602,
      "rewards/repetition_penalty_reward": -0.04219203256070614,
      "rewards/tag_count_reward": 0.9726562649011612,
      "step": 459
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 203.2083396911621,
      "epoch": 0.6899137607799025,
      "grad_norm": 125.79936841970297,
      "kl": 8.3828125,
      "learning_rate": 5.290369549525066e-06,
      "loss": 0.9341,
      "reward": 2.5255751609802246,
      "reward_std": 0.44053927063941956,
      "rewards/accuracy_reward": 0.6197916865348816,
      "rewards/reasoning_steps_reward": 0.973958358168602,
      "rewards/repetition_penalty_reward": -0.03952907770872116,
      "rewards/tag_count_reward": 0.9713541865348816,
      "step": 460
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 192.34375762939453,
      "epoch": 0.6914135733033371,
      "grad_norm": 32.0003121799561,
      "kl": 5.546875,
      "learning_rate": 5.2441680049861125e-06,
      "loss": 0.4044,
      "reward": 2.5426307916641235,
      "reward_std": 0.570788636803627,
      "rewards/accuracy_reward": 0.6979166865348816,
      "rewards/reasoning_steps_reward": 0.942708358168602,
      "rewards/repetition_penalty_reward": -0.039400530513376,
      "rewards/tag_count_reward": 0.9414062649011612,
      "step": 461
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 192.3854217529297,
      "epoch": 0.6929133858267716,
      "grad_norm": 1.8867372567328653,
      "kl": 0.93701171875,
      "learning_rate": 5.198097279683434e-06,
      "loss": 0.0048,
      "reward": 2.4953662753105164,
      "reward_std": 0.4802846685051918,
      "rewards/accuracy_reward": 0.6093750149011612,
      "rewards/reasoning_steps_reward": 0.9635416865348816,
      "rewards/repetition_penalty_reward": -0.04239422548562288,
      "rewards/tag_count_reward": 0.9648437649011612,
      "step": 462
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 218.31250381469727,
      "epoch": 0.6944131983502062,
      "grad_norm": 14.34352821328907,
      "kl": 2.4140625,
      "learning_rate": 5.152158640889947e-06,
      "loss": 0.1836,
      "reward": 2.5026406049728394,
      "reward_std": 0.5393296033143997,
      "rewards/accuracy_reward": 0.6458333432674408,
      "rewards/reasoning_steps_reward": 0.958333358168602,
      "rewards/repetition_penalty_reward": -0.05595315434038639,
      "rewards/tag_count_reward": 0.9544270932674408,
      "step": 463
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 218.9270896911621,
      "epoch": 0.6959130108736408,
      "grad_norm": 2.647430771526064,
      "kl": 0.92138671875,
      "learning_rate": 5.106353352245254e-06,
      "loss": -0.0398,
      "reward": 2.567262351512909,
      "reward_std": 0.44548412412405014,
      "rewards/accuracy_reward": 0.6979166865348816,
      "rewards/reasoning_steps_reward": 0.9635416865348816,
      "rewards/repetition_penalty_reward": -0.05252937041223049,
      "rewards/tag_count_reward": 0.958333358168602,
      "step": 464
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 231.65105056762695,
      "epoch": 0.6974128233970753,
      "grad_norm": 1.81334431827224,
      "kl": 0.81884765625,
      "learning_rate": 5.060682673720878e-06,
      "loss": -0.0188,
      "reward": 2.3878976106643677,
      "reward_std": 0.4217522442340851,
      "rewards/accuracy_reward": 0.510416679084301,
      "rewards/reasoning_steps_reward": 0.9670138955116272,
      "rewards/repetition_penalty_reward": -0.05307460017502308,
      "rewards/tag_count_reward": 0.9635416716337204,
      "step": 465
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 222.82813262939453,
      "epoch": 0.6989126359205099,
      "grad_norm": 4.851401629836836,
      "kl": 1.81005859375,
      "learning_rate": 5.015147861585603e-06,
      "loss": 0.0997,
      "reward": 2.486442983150482,
      "reward_std": 0.49670958518981934,
      "rewards/accuracy_reward": 0.5989583432674408,
      "rewards/reasoning_steps_reward": 0.9635416865348816,
      "rewards/repetition_penalty_reward": -0.03959879372268915,
      "rewards/tag_count_reward": 0.9635416716337204,
      "step": 466
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 294.14583587646484,
      "epoch": 0.7004124484439445,
      "grad_norm": 9.019683071245192,
      "kl": 1.919921875,
      "learning_rate": 4.969750168370924e-06,
      "loss": 0.2538,
      "reward": 2.627793550491333,
      "reward_std": 0.3618628829717636,
      "rewards/accuracy_reward": 0.7239583432674408,
      "rewards/reasoning_steps_reward": 0.9895833432674408,
      "rewards/repetition_penalty_reward": -0.0688210241496563,
      "rewards/tag_count_reward": 0.9830729216337204,
      "step": 467
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 248.39062881469727,
      "epoch": 0.7019122609673791,
      "grad_norm": 38.2303230093381,
      "kl": 7.177734375,
      "learning_rate": 4.924490842836584e-06,
      "loss": 0.3121,
      "reward": 2.486271381378174,
      "reward_std": 0.3947894722223282,
      "rewards/accuracy_reward": 0.5885416828095913,
      "rewards/reasoning_steps_reward": 0.9739583432674408,
      "rewards/repetition_penalty_reward": -0.04497876111418009,
      "rewards/tag_count_reward": 0.9687500149011612,
      "step": 468
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 341.1302185058594,
      "epoch": 0.7034120734908137,
      "grad_norm": 12.42233946312446,
      "kl": 1.4111328125,
      "learning_rate": 4.879371129936233e-06,
      "loss": 0.2385,
      "reward": 2.7423893213272095,
      "reward_std": 0.3288619890809059,
      "rewards/accuracy_reward": 0.848958358168602,
      "rewards/reasoning_steps_reward": 0.9947916716337204,
      "rewards/repetition_penalty_reward": -0.09094424080103636,
      "rewards/tag_count_reward": 0.9895833432674408,
      "step": 469
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 331.8645935058594,
      "epoch": 0.7049118860142483,
      "grad_norm": 4.8140326164405645,
      "kl": 3.64794921875,
      "learning_rate": 4.834392270783183e-06,
      "loss": 0.0767,
      "reward": 2.6093015670776367,
      "reward_std": 0.49695973843336105,
      "rewards/accuracy_reward": 0.7343750298023224,
      "rewards/reasoning_steps_reward": 0.9843750149011612,
      "rewards/repetition_penalty_reward": -0.08080272283405066,
      "rewards/tag_count_reward": 0.9713542014360428,
      "step": 470
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 342.6979293823242,
      "epoch": 0.7064116985376828,
      "grad_norm": 2.1505682137010256,
      "kl": 0.66015625,
      "learning_rate": 4.789555502616258e-06,
      "loss": 0.0636,
      "reward": 2.6130124926567078,
      "reward_std": 0.35396523028612137,
      "rewards/accuracy_reward": 0.7239583730697632,
      "rewards/reasoning_steps_reward": 0.9843750149011612,
      "rewards/repetition_penalty_reward": -0.07839393150061369,
      "rewards/tag_count_reward": 0.9830729365348816,
      "step": 471
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 393.6770935058594,
      "epoch": 0.7079115110611174,
      "grad_norm": 5.089873005092783,
      "kl": 0.68115234375,
      "learning_rate": 4.744862058765776e-06,
      "loss": 0.1038,
      "reward": 2.644687831401825,
      "reward_std": 0.3760165199637413,
      "rewards/accuracy_reward": 0.7812500149011612,
      "rewards/reasoning_steps_reward": 0.9947916716337204,
      "rewards/repetition_penalty_reward": -0.10661418363451958,
      "rewards/tag_count_reward": 0.9752604216337204,
      "step": 472
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 348.5104217529297,
      "epoch": 0.709411323584552,
      "grad_norm": 2.613896254667977,
      "kl": 0.80517578125,
      "learning_rate": 4.700313168619608e-06,
      "loss": 0.0745,
      "reward": 2.535249173641205,
      "reward_std": 0.3934590071439743,
      "rewards/accuracy_reward": 0.6718750298023224,
      "rewards/reasoning_steps_reward": 0.9791666865348816,
      "rewards/repetition_penalty_reward": -0.08975092135369778,
      "rewards/tag_count_reward": 0.9739583432674408,
      "step": 473
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 342.15625762939453,
      "epoch": 0.7109111361079865,
      "grad_norm": 0.927779678297728,
      "kl": 1.31298828125,
      "learning_rate": 4.655910057589377e-06,
      "loss": 0.0322,
      "reward": 2.5929868817329407,
      "reward_std": 0.37964776903390884,
      "rewards/accuracy_reward": 0.7291666716337204,
      "rewards/reasoning_steps_reward": 0.9739583432674408,
      "rewards/repetition_penalty_reward": -0.07498200424015522,
      "rewards/tag_count_reward": 0.96484375,
      "step": 474
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 286.5677185058594,
      "epoch": 0.7124109486314211,
      "grad_norm": 3.0029493567846113,
      "kl": 1.36328125,
      "learning_rate": 4.611653947076732e-06,
      "loss": 0.1304,
      "reward": 2.478628635406494,
      "reward_std": 0.5730241611599922,
      "rewards/accuracy_reward": 0.6197916865348816,
      "rewards/reasoning_steps_reward": 0.9791666865348816,
      "rewards/repetition_penalty_reward": -0.059131866320967674,
      "rewards/tag_count_reward": 0.938802108168602,
      "step": 475
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 287.01563262939453,
      "epoch": 0.7139107611548556,
      "grad_norm": 15.492730466605252,
      "kl": 4.71484375,
      "learning_rate": 4.567546054439777e-06,
      "loss": 0.2672,
      "reward": 2.366079330444336,
      "reward_std": 0.7147725075483322,
      "rewards/accuracy_reward": 0.5885416865348816,
      "rewards/reasoning_steps_reward": 0.9756944626569748,
      "rewards/repetition_penalty_reward": -0.06143816187977791,
      "rewards/tag_count_reward": 0.8632812649011612,
      "step": 476
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 298.7343864440918,
      "epoch": 0.7154105736782902,
      "grad_norm": 49.985480375628754,
      "kl": 8.77734375,
      "learning_rate": 4.523587592959557e-06,
      "loss": 0.6553,
      "reward": 2.199244737625122,
      "reward_std": 0.696681559085846,
      "rewards/accuracy_reward": 0.463541679084301,
      "rewards/reasoning_steps_reward": 0.9704861044883728,
      "rewards/repetition_penalty_reward": -0.07853319868445396,
      "rewards/tag_count_reward": 0.8437500298023224,
      "step": 477
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 264.0104293823242,
      "epoch": 0.7169103862017248,
      "grad_norm": 9.130997990308742,
      "kl": 2.22802734375,
      "learning_rate": 4.479779771806699e-06,
      "loss": 0.2573,
      "reward": 2.621893048286438,
      "reward_std": 0.5179613158106804,
      "rewards/accuracy_reward": 0.7291666865348816,
      "rewards/reasoning_steps_reward": 0.9947916716337204,
      "rewards/repetition_penalty_reward": -0.05779455881565809,
      "rewards/tag_count_reward": 0.9557292014360428,
      "step": 478
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 250.98958587646484,
      "epoch": 0.7184101987251593,
      "grad_norm": 1.642461450377282,
      "kl": 1.078125,
      "learning_rate": 4.436123796008149e-06,
      "loss": 0.081,
      "reward": 2.540480315685272,
      "reward_std": 0.4395308271050453,
      "rewards/accuracy_reward": 0.6406250149011612,
      "rewards/reasoning_steps_reward": 0.9791666716337204,
      "rewards/repetition_penalty_reward": -0.046759432181715965,
      "rewards/tag_count_reward": 0.9674479216337204,
      "step": 479
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 264.43750762939453,
      "epoch": 0.7199100112485939,
      "grad_norm": 2.118847794368944,
      "kl": 0.8173828125,
      "learning_rate": 4.392620866414026e-06,
      "loss": 0.0134,
      "reward": 2.5877268314361572,
      "reward_std": 0.45883308351039886,
      "rewards/accuracy_reward": 0.697916679084301,
      "rewards/reasoning_steps_reward": 0.9826389104127884,
      "rewards/repetition_penalty_reward": -0.05246431287378073,
      "rewards/tag_count_reward": 0.9596354365348816,
      "step": 480
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 263.8698043823242,
      "epoch": 0.7214098237720284,
      "grad_norm": 2.650392492259305,
      "kl": 0.63818359375,
      "learning_rate": 4.349272179664586e-06,
      "loss": 0.0475,
      "reward": 2.623681426048279,
      "reward_std": 0.4385247528553009,
      "rewards/accuracy_reward": 0.7083333432674408,
      "rewards/reasoning_steps_reward": 0.9843750149011612,
      "rewards/repetition_penalty_reward": -0.04689165949821472,
      "rewards/tag_count_reward": 0.977864608168602,
      "step": 481
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 260.54687881469727,
      "epoch": 0.7229096362954631,
      "grad_norm": 1.8488154063721305,
      "kl": 1.2900390625,
      "learning_rate": 4.3060789281573135e-06,
      "loss": 0.0759,
      "reward": 2.620662033557892,
      "reward_std": 0.5110099017620087,
      "rewards/accuracy_reward": 0.755208358168602,
      "rewards/reasoning_steps_reward": 0.9635416716337204,
      "rewards/repetition_penalty_reward": -0.053817191161215305,
      "rewards/tag_count_reward": 0.9557291716337204,
      "step": 482
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 236.7708396911621,
      "epoch": 0.7244094488188977,
      "grad_norm": 0.8685220289150081,
      "kl": 0.9541015625,
      "learning_rate": 4.263042300014112e-06,
      "loss": 0.0159,
      "reward": 2.6224602460861206,
      "reward_std": 0.432863712310791,
      "rewards/accuracy_reward": 0.7135416865348816,
      "rewards/reasoning_steps_reward": 0.975694477558136,
      "rewards/repetition_penalty_reward": -0.036828051786869764,
      "rewards/tag_count_reward": 0.9700520932674408,
      "step": 483
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 247.7604217529297,
      "epoch": 0.7259092613423322,
      "grad_norm": 8.634827245487672,
      "kl": 2.59814453125,
      "learning_rate": 4.220163479048632e-06,
      "loss": 0.0962,
      "reward": 2.589983582496643,
      "reward_std": 0.55703204870224,
      "rewards/accuracy_reward": 0.7239583507180214,
      "rewards/reasoning_steps_reward": 0.9548611491918564,
      "rewards/repetition_penalty_reward": -0.039356810972094536,
      "rewards/tag_count_reward": 0.9505208432674408,
      "step": 484
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 243.9947967529297,
      "epoch": 0.7274090738657668,
      "grad_norm": 2.8247050710119654,
      "kl": 1.40234375,
      "learning_rate": 4.177443644733699e-06,
      "loss": 0.0324,
      "reward": 2.7192403078079224,
      "reward_std": 0.2972189523279667,
      "rewards/accuracy_reward": 0.8020833432674408,
      "rewards/reasoning_steps_reward": 0.984375,
      "rewards/repetition_penalty_reward": -0.04117643600329757,
      "rewards/tag_count_reward": 0.973958358168602,
      "step": 485
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 251.07291793823242,
      "epoch": 0.7289088863892014,
      "grad_norm": 24.101080804460526,
      "kl": 4.90234375,
      "learning_rate": 4.134883972168877e-06,
      "loss": 0.3001,
      "reward": 2.681569218635559,
      "reward_std": 0.5304646193981171,
      "rewards/accuracy_reward": 0.802083358168602,
      "rewards/reasoning_steps_reward": 0.9618055671453476,
      "rewards/repetition_penalty_reward": -0.03935106098651886,
      "rewards/tag_count_reward": 0.9570312649011612,
      "step": 486
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 249.77604293823242,
      "epoch": 0.7304086989126359,
      "grad_norm": 9.052170965793191,
      "kl": 2.7578125,
      "learning_rate": 4.092485632048142e-06,
      "loss": 0.2795,
      "reward": 2.5998495221138,
      "reward_std": 0.3634401187300682,
      "rewards/accuracy_reward": 0.6718750074505806,
      "rewards/reasoning_steps_reward": 0.9843750149011612,
      "rewards/repetition_penalty_reward": -0.031660950277000666,
      "rewards/tag_count_reward": 0.9752604365348816,
      "step": 487
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 257.9947967529297,
      "epoch": 0.7319085114360705,
      "grad_norm": 1.4255147156673067,
      "kl": 1.27685546875,
      "learning_rate": 4.050249790627675e-06,
      "loss": 0.1139,
      "reward": 2.5903435349464417,
      "reward_std": 0.47754330188035965,
      "rewards/accuracy_reward": 0.677083358168602,
      "rewards/reasoning_steps_reward": 0.9809027910232544,
      "rewards/repetition_penalty_reward": -0.04029886703938246,
      "rewards/tag_count_reward": 0.9726562649011612,
      "step": 488
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 243.2864646911621,
      "epoch": 0.7334083239595051,
      "grad_norm": 8.575587858559526,
      "kl": 3.0546875,
      "learning_rate": 4.008177609693791e-06,
      "loss": 0.0417,
      "reward": 2.663625717163086,
      "reward_std": 0.5341070555150509,
      "rewards/accuracy_reward": 0.7812500149011612,
      "rewards/reasoning_steps_reward": 0.960069477558136,
      "rewards/repetition_penalty_reward": -0.03342306334525347,
      "rewards/tag_count_reward": 0.9557292014360428,
      "step": 489
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 245.7395896911621,
      "epoch": 0.7349081364829396,
      "grad_norm": 2.765671602090238,
      "kl": 0.9609375,
      "learning_rate": 3.966270246530975e-06,
      "loss": 0.0215,
      "reward": 2.6076095700263977,
      "reward_std": 0.4131758604198694,
      "rewards/accuracy_reward": 0.692708358168602,
      "rewards/reasoning_steps_reward": 0.9791666716337204,
      "rewards/repetition_penalty_reward": -0.03561965608969331,
      "rewards/tag_count_reward": 0.9713541865348816,
      "step": 490
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 277.2135543823242,
      "epoch": 0.7364079490063742,
      "grad_norm": 2.188660242549928,
      "kl": 0.86328125,
      "learning_rate": 3.924528853890046e-06,
      "loss": 0.0266,
      "reward": 2.5408560633659363,
      "reward_std": 0.4740516468882561,
      "rewards/accuracy_reward": 0.6718750223517418,
      "rewards/reasoning_steps_reward": 0.9687500149011612,
      "rewards/repetition_penalty_reward": -0.05159200169146061,
      "rewards/tag_count_reward": 0.9518229365348816,
      "step": 491
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 270.3020935058594,
      "epoch": 0.7379077615298087,
      "grad_norm": 3.1759937413803816,
      "kl": 0.900390625,
      "learning_rate": 3.882954579956455e-06,
      "loss": 0.001,
      "reward": 2.583053767681122,
      "reward_std": 0.4794854000210762,
      "rewards/accuracy_reward": 0.6875000149011612,
      "rewards/reasoning_steps_reward": 0.973958358168602,
      "rewards/repetition_penalty_reward": -0.043248409405350685,
      "rewards/tag_count_reward": 0.9648437798023224,
      "step": 492
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 282.2239761352539,
      "epoch": 0.7394075740532433,
      "grad_norm": 2.975773686014791,
      "kl": 1.705078125,
      "learning_rate": 3.841548568318706e-06,
      "loss": 0.0655,
      "reward": 2.4317139387130737,
      "reward_std": 0.6113808900117874,
      "rewards/accuracy_reward": 0.5885416716337204,
      "rewards/reasoning_steps_reward": 0.9565972536802292,
      "rewards/repetition_penalty_reward": -0.048320941627025604,
      "rewards/tag_count_reward": 0.9348958432674408,
      "step": 493
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 249.14583587646484,
      "epoch": 0.7409073865766779,
      "grad_norm": 1.441571062260193,
      "kl": 1.27587890625,
      "learning_rate": 3.8003119579368806e-06,
      "loss": 0.0064,
      "reward": 2.4567288756370544,
      "reward_std": 0.39962563663721085,
      "rewards/accuracy_reward": 0.541666679084301,
      "rewards/reasoning_steps_reward": 0.9739583432674408,
      "rewards/repetition_penalty_reward": -0.03285460267215967,
      "rewards/tag_count_reward": 0.9739583432674408,
      "step": 494
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 259.42188262939453,
      "epoch": 0.7424071991001124,
      "grad_norm": 2.6425323421375118,
      "kl": 0.888671875,
      "learning_rate": 3.7592458831113256e-06,
      "loss": 0.0337,
      "reward": 2.3374595046043396,
      "reward_std": 0.3100433573126793,
      "rewards/accuracy_reward": 0.38541668839752674,
      "rewards/reasoning_steps_reward": 1.0,
      "rewards/repetition_penalty_reward": -0.04144685994833708,
      "rewards/tag_count_reward": 0.9934895932674408,
      "step": 495
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 285.37500762939453,
      "epoch": 0.7439070116235471,
      "grad_norm": 1.6657973548432092,
      "kl": 1.005859375,
      "learning_rate": 3.718351473451448e-06,
      "loss": 0.0419,
      "reward": 2.6762454509735107,
      "reward_std": 0.2155936686322093,
      "rewards/accuracy_reward": 0.7500000149011612,
      "rewards/reasoning_steps_reward": 0.9895833432674408,
      "rewards/repetition_penalty_reward": -0.04380669817328453,
      "rewards/tag_count_reward": 0.9804687649011612,
      "step": 496
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 288.8593864440918,
      "epoch": 0.7454068241469817,
      "grad_norm": 0.7716130912654924,
      "kl": 0.6650390625,
      "learning_rate": 3.6776298538446307e-06,
      "loss": 0.0417,
      "reward": 2.51417738199234,
      "reward_std": 0.32952259480953217,
      "rewards/accuracy_reward": 0.5989583432674408,
      "rewards/reasoning_steps_reward": 0.9843750149011612,
      "rewards/repetition_penalty_reward": -0.04962483886629343,
      "rewards/tag_count_reward": 0.9804687798023224,
      "step": 497
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 310.2760543823242,
      "epoch": 0.7469066366704162,
      "grad_norm": 5.134089109133943,
      "kl": 2.931640625,
      "learning_rate": 3.6370821444253112e-06,
      "loss": 0.0652,
      "reward": 2.4335711002349854,
      "reward_std": 0.3948415219783783,
      "rewards/accuracy_reward": 0.5572916865348816,
      "rewards/reasoning_steps_reward": 0.9687500149011612,
      "rewards/repetition_penalty_reward": -0.050803929567337036,
      "rewards/tag_count_reward": 0.958333358168602,
      "step": 498
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 295.4270935058594,
      "epoch": 0.7484064491938508,
      "grad_norm": 2.574704298931235,
      "kl": 1.828125,
      "learning_rate": 3.5967094605441545e-06,
      "loss": -0.0396,
      "reward": 2.6697959899902344,
      "reward_std": 0.49928344041109085,
      "rewards/accuracy_reward": 0.7968750149011612,
      "rewards/reasoning_steps_reward": 0.9670139253139496,
      "rewards/repetition_penalty_reward": -0.05112431012094021,
      "rewards/tag_count_reward": 0.9570312649011612,
      "step": 499
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 264.76563262939453,
      "epoch": 0.7499062617172854,
      "grad_norm": 1.121236474277334,
      "kl": 0.8974609375,
      "learning_rate": 3.5565129127373765e-06,
      "loss": -0.0091,
      "reward": 2.7433581352233887,
      "reward_std": 0.39212197065353394,
      "rewards/accuracy_reward": 0.8229166865348816,
      "rewards/reasoning_steps_reward": 0.9843750149011612,
      "rewards/repetition_penalty_reward": -0.04049617797136307,
      "rewards/tag_count_reward": 0.9765625149011612,
      "step": 500
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 271.3906364440918,
      "epoch": 0.7514060742407199,
      "grad_norm": 0.7018570637816092,
      "kl": 1.29638671875,
      "learning_rate": 3.5164936066961984e-06,
      "loss": -0.0223,
      "reward": 2.6698635816574097,
      "reward_std": 0.3935448888223618,
      "rewards/accuracy_reward": 0.7760416865348816,
      "rewards/reasoning_steps_reward": 0.9687500149011612,
      "rewards/repetition_penalty_reward": -0.03846996137872338,
      "rewards/tag_count_reward": 0.9635416865348816,
      "step": 501
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 314.0052185058594,
      "epoch": 0.7529058867641545,
      "grad_norm": 1.1907324808542725,
      "kl": 0.833984375,
      "learning_rate": 3.476652643236431e-06,
      "loss": 0.0334,
      "reward": 2.6077919006347656,
      "reward_std": 0.3832826167345047,
      "rewards/accuracy_reward": 0.7291666865348816,
      "rewards/reasoning_steps_reward": 0.9739583432674408,
      "rewards/repetition_penalty_reward": -0.05757272336632013,
      "rewards/tag_count_reward": 0.962239608168602,
      "step": 502
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 317.2448043823242,
      "epoch": 0.754405699287589,
      "grad_norm": 1.0569480071148656,
      "kl": 1.544921875,
      "learning_rate": 3.436991118268195e-06,
      "loss": 0.0521,
      "reward": 2.522133708000183,
      "reward_std": 0.4893868714570999,
      "rewards/accuracy_reward": 0.6562500223517418,
      "rewards/reasoning_steps_reward": 0.973958358168602,
      "rewards/repetition_penalty_reward": -0.06119971442967653,
      "rewards/tag_count_reward": 0.9531250298023224,
      "step": 503
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 301.3802146911621,
      "epoch": 0.7559055118110236,
      "grad_norm": 5.910389855911775,
      "kl": 11.95703125,
      "learning_rate": 3.3975101227657726e-06,
      "loss": -0.1243,
      "reward": 2.3981975317001343,
      "reward_std": 0.6437404751777649,
      "rewards/accuracy_reward": 0.5468750223517418,
      "rewards/reasoning_steps_reward": 0.958333358168602,
      "rewards/repetition_penalty_reward": -0.05102120712399483,
      "rewards/tag_count_reward": 0.9440104514360428,
      "step": 504
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 340.42189025878906,
      "epoch": 0.7574053243344582,
      "grad_norm": 12.978062482129692,
      "kl": 16.43701171875,
      "learning_rate": 3.3582107427376044e-06,
      "loss": -0.0418,
      "reward": 2.647988438606262,
      "reward_std": 0.5043715462088585,
      "rewards/accuracy_reward": 0.770833358168602,
      "rewards/reasoning_steps_reward": 0.9739583432674408,
      "rewards/repetition_penalty_reward": -0.057740709744393826,
      "rewards/tag_count_reward": 0.9609375149011612,
      "step": 505
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 349.92188262939453,
      "epoch": 0.7589051368578927,
      "grad_norm": 6.073801480071967,
      "kl": 1.833984375,
      "learning_rate": 3.3190940591964094e-06,
      "loss": 0.0047,
      "reward": 2.440945327281952,
      "reward_std": 0.4255194067955017,
      "rewards/accuracy_reward": 0.5781250149011612,
      "rewards/reasoning_steps_reward": 0.972222238779068,
      "rewards/repetition_penalty_reward": -0.07554787117987871,
      "rewards/tag_count_reward": 0.966145858168602,
      "step": 506
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 413.4791793823242,
      "epoch": 0.7604049493813273,
      "grad_norm": 2.37952801915241,
      "kl": 4.4140625,
      "learning_rate": 3.2801611481294538e-06,
      "loss": 0.0238,
      "reward": 2.686852276325226,
      "reward_std": 0.3788864966481924,
      "rewards/accuracy_reward": 0.8385416716337204,
      "rewards/reasoning_steps_reward": 0.9774305820465088,
      "rewards/repetition_penalty_reward": -0.08875534310936928,
      "rewards/tag_count_reward": 0.9596354365348816,
      "step": 507
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 381.5520935058594,
      "epoch": 0.7619047619047619,
      "grad_norm": 0.9844048707484628,
      "kl": 1.52587890625,
      "learning_rate": 3.2414130804689492e-06,
      "loss": 0.0616,
      "reward": 2.4849266409873962,
      "reward_std": 0.44034913927316666,
      "rewards/accuracy_reward": 0.6093750074505806,
      "rewards/reasoning_steps_reward": 0.9843750149011612,
      "rewards/repetition_penalty_reward": -0.06845879275351763,
      "rewards/tag_count_reward": 0.9596354365348816,
      "step": 508
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 337.81251525878906,
      "epoch": 0.7634045744281964,
      "grad_norm": 0.8845374524712641,
      "kl": 1.2822265625,
      "learning_rate": 3.202850922062607e-06,
      "loss": -0.0007,
      "reward": 2.6837249398231506,
      "reward_std": 0.5106394588947296,
      "rewards/accuracy_reward": 0.8125000298023224,
      "rewards/reasoning_steps_reward": 0.973958358168602,
      "rewards/repetition_penalty_reward": -0.06106678955256939,
      "rewards/tag_count_reward": 0.958333358168602,
      "step": 509
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 366.96875762939453,
      "epoch": 0.7649043869516311,
      "grad_norm": 0.6486204612451865,
      "kl": 1.52978515625,
      "learning_rate": 3.1644757336443023e-06,
      "loss": -0.0686,
      "reward": 2.4383880496025085,
      "reward_std": 0.602533221244812,
      "rewards/accuracy_reward": 0.609375,
      "rewards/reasoning_steps_reward": 0.958333358168602,
      "rewards/repetition_penalty_reward": -0.06551818642765284,
      "rewards/tag_count_reward": 0.9361979365348816,
      "step": 510
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 355.81250762939453,
      "epoch": 0.7664041994750657,
      "grad_norm": 0.8344794010535831,
      "kl": 0.908203125,
      "learning_rate": 3.126288570804906e-06,
      "loss": 0.0427,
      "reward": 2.516266345977783,
      "reward_std": 0.5227550566196442,
      "rewards/accuracy_reward": 0.635416679084301,
      "rewards/reasoning_steps_reward": 0.9826388955116272,
      "rewards/repetition_penalty_reward": -0.05882055405527353,
      "rewards/tag_count_reward": 0.9570312649011612,
      "step": 511
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 365.3385467529297,
      "epoch": 0.7679040119985002,
      "grad_norm": 159.77587834072062,
      "kl": 10.79296875,
      "learning_rate": 3.0882904839632476e-06,
      "loss": 0.6306,
      "reward": 2.5336210131645203,
      "reward_std": 0.548546776175499,
      "rewards/accuracy_reward": 0.6718750149011612,
      "rewards/reasoning_steps_reward": 0.9791666865348816,
      "rewards/repetition_penalty_reward": -0.06143125705420971,
      "rewards/tag_count_reward": 0.9440104365348816,
      "step": 512
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 364.0677185058594,
      "epoch": 0.7694038245219348,
      "grad_norm": 45.25299778164268,
      "kl": 2.84375,
      "learning_rate": 3.050482518337221e-06,
      "loss": 0.1528,
      "reward": 2.5458881855010986,
      "reward_std": 0.5097367987036705,
      "rewards/accuracy_reward": 0.6927083432674408,
      "rewards/reasoning_steps_reward": 0.9687500298023224,
      "rewards/repetition_penalty_reward": -0.06739312317222357,
      "rewards/tag_count_reward": 0.9518229365348816,
      "step": 513
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 346.5416793823242,
      "epoch": 0.7709036370453693,
      "grad_norm": 1.324438735835731,
      "kl": 1.3828125,
      "learning_rate": 3.012865713915033e-06,
      "loss": -0.0002,
      "reward": 2.548605740070343,
      "reward_std": 0.4542975649237633,
      "rewards/accuracy_reward": 0.6822916865348816,
      "rewards/reasoning_steps_reward": 0.9791666865348816,
      "rewards/repetition_penalty_reward": -0.05686304159462452,
      "rewards/tag_count_reward": 0.9440104365348816,
      "step": 514
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 372.2552261352539,
      "epoch": 0.7724034495688039,
      "grad_norm": 1.3346884321616983,
      "kl": 1.03759765625,
      "learning_rate": 2.9754411054265966e-06,
      "loss": 0.033,
      "reward": 2.5564919114112854,
      "reward_std": 0.36123840510845184,
      "rewards/accuracy_reward": 0.6562500149011612,
      "rewards/reasoning_steps_reward": 0.9895833432674408,
      "rewards/repetition_penalty_reward": -0.05809137877076864,
      "rewards/tag_count_reward": 0.9687500298023224,
      "step": 515
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 350.9791717529297,
      "epoch": 0.7739032620922385,
      "grad_norm": 4.616902608747506,
      "kl": 3.259765625,
      "learning_rate": 2.9382097223150675e-06,
      "loss": 0.0779,
      "reward": 2.524458348751068,
      "reward_std": 0.5577153712511063,
      "rewards/accuracy_reward": 0.677083358168602,
      "rewards/reasoning_steps_reward": 0.9687500149011612,
      "rewards/repetition_penalty_reward": -0.0549686960875988,
      "rewards/tag_count_reward": 0.9335937649011612,
      "step": 516
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 437.5729217529297,
      "epoch": 0.775403074615673,
      "grad_norm": 2.0004592558673195,
      "kl": 0.81689453125,
      "learning_rate": 2.9011725887085286e-06,
      "loss": 0.0369,
      "reward": 2.481614410877228,
      "reward_std": 0.5043549910187721,
      "rewards/accuracy_reward": 0.6093750298023224,
      "rewards/reasoning_steps_reward": 0.9843750149011612,
      "rewards/repetition_penalty_reward": -0.07177106104791164,
      "rewards/tag_count_reward": 0.9596354365348816,
      "step": 517
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 333.14063262939453,
      "epoch": 0.7769028871391076,
      "grad_norm": 2.055433226987207,
      "kl": 2.642578125,
      "learning_rate": 2.8643307233918192e-06,
      "loss": 0.0882,
      "reward": 2.5522547364234924,
      "reward_std": 0.5509222820401192,
      "rewards/accuracy_reward": 0.708333358168602,
      "rewards/reasoning_steps_reward": 0.9618055820465088,
      "rewards/repetition_penalty_reward": -0.05408210679888725,
      "rewards/tag_count_reward": 0.9361979365348816,
      "step": 518
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 391.0208435058594,
      "epoch": 0.7784026996625422,
      "grad_norm": 4.053898086811571,
      "kl": 2.04736328125,
      "learning_rate": 2.827685139778511e-06,
      "loss": 0.0504,
      "reward": 2.5092581510543823,
      "reward_std": 0.5159156918525696,
      "rewards/accuracy_reward": 0.6354166716337204,
      "rewards/reasoning_steps_reward": 0.9861111342906952,
      "rewards/repetition_penalty_reward": -0.06148845702409744,
      "rewards/tag_count_reward": 0.9492187649011612,
      "step": 519
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 337.0885543823242,
      "epoch": 0.7799025121859767,
      "grad_norm": 0.8528892895209994,
      "kl": 1.2861328125,
      "learning_rate": 2.7912368458830295e-06,
      "loss": 0.0657,
      "reward": 2.51828670501709,
      "reward_std": 0.46771006286144257,
      "rewards/accuracy_reward": 0.6250000149011612,
      "rewards/reasoning_steps_reward": 0.9843750149011612,
      "rewards/repetition_penalty_reward": -0.05463008023798466,
      "rewards/tag_count_reward": 0.9635416865348816,
      "step": 520
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 324.42708587646484,
      "epoch": 0.7814023247094113,
      "grad_norm": 1.5995624359002987,
      "kl": 1.50927734375,
      "learning_rate": 2.7549868442929286e-06,
      "loss": 0.0114,
      "reward": 2.534961521625519,
      "reward_std": 0.42034848034381866,
      "rewards/accuracy_reward": 0.6510416716337204,
      "rewards/reasoning_steps_reward": 0.9843750149011612,
      "rewards/repetition_penalty_reward": -0.04706980008631945,
      "rewards/tag_count_reward": 0.946614608168602,
      "step": 521
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 343.7395935058594,
      "epoch": 0.7829021372328459,
      "grad_norm": 2.8590022801404706,
      "kl": 1.525390625,
      "learning_rate": 2.7189361321413144e-06,
      "loss": 0.1087,
      "reward": 2.500767946243286,
      "reward_std": 0.4306958243250847,
      "rewards/accuracy_reward": 0.6041666865348816,
      "rewards/reasoning_steps_reward": 0.9809028059244156,
      "rewards/repetition_penalty_reward": -0.05044758692383766,
      "rewards/tag_count_reward": 0.9661458432674408,
      "step": 522
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 353.6145935058594,
      "epoch": 0.7844019497562804,
      "grad_norm": 0.9462589535923942,
      "kl": 1.6015625,
      "learning_rate": 2.683085701079412e-06,
      "loss": 0.0087,
      "reward": 2.624489724636078,
      "reward_std": 0.46847573667764664,
      "rewards/accuracy_reward": 0.7447916865348816,
      "rewards/reasoning_steps_reward": 0.9809028059244156,
      "rewards/repetition_penalty_reward": -0.05563185177743435,
      "rewards/tag_count_reward": 0.9544270932674408,
      "step": 523
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 277.2135467529297,
      "epoch": 0.7859017622797151,
      "grad_norm": 0.6580260031686143,
      "kl": 1.3935546875,
      "learning_rate": 2.647436537249294e-06,
      "loss": -0.0381,
      "reward": 2.4149619936943054,
      "reward_std": 0.39340294152498245,
      "rewards/accuracy_reward": 0.5052083544433117,
      "rewards/reasoning_steps_reward": 0.9791666865348816,
      "rewards/repetition_penalty_reward": -0.036860992200672626,
      "rewards/tag_count_reward": 0.9674479365348816,
      "step": 524
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 323.4114761352539,
      "epoch": 0.7874015748031497,
      "grad_norm": 1.021591339198619,
      "kl": 1.7890625,
      "learning_rate": 2.611989621256745e-06,
      "loss": 0.0141,
      "reward": 2.434917449951172,
      "reward_std": 0.5576092228293419,
      "rewards/accuracy_reward": 0.588541679084301,
      "rewards/reasoning_steps_reward": 0.9583333730697632,
      "rewards/repetition_penalty_reward": -0.05075975460931659,
      "rewards/tag_count_reward": 0.938802108168602,
      "step": 525
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 251.5000114440918,
      "epoch": 0.7889013873265842,
      "grad_norm": 2.0873915039216073,
      "kl": 2.01171875,
      "learning_rate": 2.5767459281443064e-06,
      "loss": -0.0672,
      "reward": 2.2912232875823975,
      "reward_std": 0.42987949773669243,
      "rewards/accuracy_reward": 0.42708335584029555,
      "rewards/reasoning_steps_reward": 0.9652778059244156,
      "rewards/repetition_penalty_reward": -0.036033790558576584,
      "rewards/tag_count_reward": 0.9348958432674408,
      "step": 526
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 308.4895896911621,
      "epoch": 0.7904011998500188,
      "grad_norm": 1.7291159865969803,
      "kl": 1.93310546875,
      "learning_rate": 2.541706427364431e-06,
      "loss": -0.0153,
      "reward": 2.6689711213111877,
      "reward_std": 0.5232843309640884,
      "rewards/accuracy_reward": 0.7760416716337204,
      "rewards/reasoning_steps_reward": 0.9809028059244156,
      "rewards/repetition_penalty_reward": -0.04500464163720608,
      "rewards/tag_count_reward": 0.9570312798023224,
      "step": 527
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 264.5572967529297,
      "epoch": 0.7919010123734533,
      "grad_norm": 1.784246342813351,
      "kl": 0.90576171875,
      "learning_rate": 2.506872082752834e-06,
      "loss": 0.0971,
      "reward": 2.664152204990387,
      "reward_std": 0.30452917888760567,
      "rewards/accuracy_reward": 0.7239583432674408,
      "rewards/reasoning_steps_reward": 0.9947916716337204,
      "rewards/repetition_penalty_reward": -0.03376449551433325,
      "rewards/tag_count_reward": 0.9791666865348816,
      "step": 528
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 325.4791793823242,
      "epoch": 0.7934008248968879,
      "grad_norm": 3.266992308801891,
      "kl": 1.8212890625,
      "learning_rate": 2.4722438525019764e-06,
      "loss": 0.0034,
      "reward": 2.676040768623352,
      "reward_std": 0.44805118814110756,
      "rewards/accuracy_reward": 0.770833358168602,
      "rewards/reasoning_steps_reward": 0.9878472238779068,
      "rewards/repetition_penalty_reward": -0.05399402230978012,
      "rewards/tag_count_reward": 0.9713541865348816,
      "step": 529
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 258.14063262939453,
      "epoch": 0.7949006374203225,
      "grad_norm": 2.2367523347428366,
      "kl": 2.3408203125,
      "learning_rate": 2.4378226891347056e-06,
      "loss": -0.0131,
      "reward": 2.6123253107070923,
      "reward_std": 0.4946172907948494,
      "rewards/accuracy_reward": 0.7291666865348816,
      "rewards/reasoning_steps_reward": 0.9687500149011612,
      "rewards/repetition_penalty_reward": -0.03611219022423029,
      "rewards/tag_count_reward": 0.9505208432674408,
      "step": 530
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 357.5104293823242,
      "epoch": 0.796400449943757,
      "grad_norm": 2.0536966546593494,
      "kl": 2.0966796875,
      "learning_rate": 2.403609539478056e-06,
      "loss": 0.015,
      "reward": 2.6887767910957336,
      "reward_std": 0.43632883578538895,
      "rewards/accuracy_reward": 0.8072917014360428,
      "rewards/reasoning_steps_reward": 0.9895833432674408,
      "rewards/repetition_penalty_reward": -0.06382743921130896,
      "rewards/tag_count_reward": 0.9557291716337204,
      "step": 531
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 291.60937881469727,
      "epoch": 0.7979002624671916,
      "grad_norm": 68.06193885878405,
      "kl": 3.470703125,
      "learning_rate": 2.3696053446372026e-06,
      "loss": 0.1695,
      "reward": 2.4727484583854675,
      "reward_std": 0.5058952420949936,
      "rewards/accuracy_reward": 0.5833333432674408,
      "rewards/reasoning_steps_reward": 0.9809028059244156,
      "rewards/repetition_penalty_reward": -0.04721691645681858,
      "rewards/tag_count_reward": 0.9557291865348816,
      "step": 532
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 283.70833587646484,
      "epoch": 0.7994000749906262,
      "grad_norm": 7.175346457591991,
      "kl": 2.5625,
      "learning_rate": 2.3358110399695788e-06,
      "loss": 0.1764,
      "reward": 2.499284565448761,
      "reward_std": 0.42855924367904663,
      "rewards/accuracy_reward": 0.5989583507180214,
      "rewards/reasoning_steps_reward": 0.9791666716337204,
      "rewards/repetition_penalty_reward": -0.04108010325580835,
      "rewards/tag_count_reward": 0.9622396230697632,
      "step": 533
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 247.94792556762695,
      "epoch": 0.8008998875140607,
      "grad_norm": 1.6428729718898072,
      "kl": 0.5927734375,
      "learning_rate": 2.302227555059141e-06,
      "loss": 0.0496,
      "reward": 2.5920801758766174,
      "reward_std": 0.3266643173992634,
      "rewards/accuracy_reward": 0.6458333432674408,
      "rewards/reasoning_steps_reward": 1.0,
      "rewards/repetition_penalty_reward": -0.03422196349129081,
      "rewards/tag_count_reward": 0.9804687798023224,
      "step": 534
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 309.5833396911621,
      "epoch": 0.8023997000374953,
      "grad_norm": 47.540127693483335,
      "kl": 4.4296875,
      "learning_rate": 2.2688558136908025e-06,
      "loss": 0.0615,
      "reward": 2.5951295495033264,
      "reward_std": 0.41650524735450745,
      "rewards/accuracy_reward": 0.6822916865348816,
      "rewards/reasoning_steps_reward": 0.9895833432674408,
      "rewards/repetition_penalty_reward": -0.04940184485167265,
      "rewards/tag_count_reward": 0.9726562649011612,
      "step": 535
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 260.6510467529297,
      "epoch": 0.8038995125609298,
      "grad_norm": 15.375690774431247,
      "kl": 17.03125,
      "learning_rate": 2.2356967338250223e-06,
      "loss": 0.1811,
      "reward": 2.492598533630371,
      "reward_std": 0.4195599779486656,
      "rewards/accuracy_reward": 0.5989583507180214,
      "rewards/reasoning_steps_reward": 0.973958358168602,
      "rewards/repetition_penalty_reward": -0.03604749217629433,
      "rewards/tag_count_reward": 0.9557291865348816,
      "step": 536
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 292.3854217529297,
      "epoch": 0.8053993250843644,
      "grad_norm": 9.475530371016637,
      "kl": 2.486328125,
      "learning_rate": 2.202751227572556e-06,
      "loss": 0.0533,
      "reward": 2.648472845554352,
      "reward_std": 0.3912891000509262,
      "rewards/accuracy_reward": 0.7343750298023224,
      "rewards/reasoning_steps_reward": 0.9895833432674408,
      "rewards/repetition_penalty_reward": -0.050745952408760786,
      "rewards/tag_count_reward": 0.9752604365348816,
      "step": 537
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 315.9739761352539,
      "epoch": 0.8068991376077991,
      "grad_norm": 1.4250516842234306,
      "kl": 1.0771484375,
      "learning_rate": 2.1700202011693573e-06,
      "loss": 0.0594,
      "reward": 2.7068479657173157,
      "reward_std": 0.3395584188401699,
      "rewards/accuracy_reward": 0.7968750149011612,
      "rewards/reasoning_steps_reward": 0.9895833432674408,
      "rewards/repetition_penalty_reward": -0.05487091187387705,
      "rewards/tag_count_reward": 0.9752604514360428,
      "step": 538
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 266.8489646911621,
      "epoch": 0.8083989501312336,
      "grad_norm": 3.4214757569933,
      "kl": 1.81640625,
      "learning_rate": 2.1375045549516636e-06,
      "loss": 0.0096,
      "reward": 2.5113611817359924,
      "reward_std": 0.5310123562812805,
      "rewards/accuracy_reward": 0.6250000149011612,
      "rewards/reasoning_steps_reward": 0.973958358168602,
      "rewards/repetition_penalty_reward": -0.044628492556512356,
      "rewards/tag_count_reward": 0.9570312649011612,
      "step": 539
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 255.57292556762695,
      "epoch": 0.8098987626546682,
      "grad_norm": 2.1241861424196147,
      "kl": 1.56494140625,
      "learning_rate": 2.105205183331224e-06,
      "loss": 0.0114,
      "reward": 2.7599127888679504,
      "reward_std": 0.18678605183959007,
      "rewards/accuracy_reward": 0.8333333432674408,
      "rewards/reasoning_steps_reward": 0.9895833432674408,
      "rewards/repetition_penalty_reward": -0.04477462079375982,
      "rewards/tag_count_reward": 0.9817708432674408,
      "step": 540
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 331.3385543823242,
      "epoch": 0.8113985751781028,
      "grad_norm": 1.5714564014928512,
      "kl": 1.33203125,
      "learning_rate": 2.0731229747706926e-06,
      "loss": 0.0368,
      "reward": 2.4418932795524597,
      "reward_std": 0.4528404325246811,
      "rewards/accuracy_reward": 0.5468750149011612,
      "rewards/reasoning_steps_reward": 0.9895833432674408,
      "rewards/repetition_penalty_reward": -0.058106820564717054,
      "rewards/tag_count_reward": 0.9635416865348816,
      "step": 541
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 344.2760467529297,
      "epoch": 0.8128983877015373,
      "grad_norm": 0.7740942337468878,
      "kl": 1.365234375,
      "learning_rate": 2.041258811759195e-06,
      "loss": -0.0408,
      "reward": 2.485717535018921,
      "reward_std": 0.3474634326994419,
      "rewards/accuracy_reward": 0.5937500111758709,
      "rewards/reasoning_steps_reward": 0.9895833432674408,
      "rewards/repetition_penalty_reward": -0.07417836599051952,
      "rewards/tag_count_reward": 0.9765625149011612,
      "step": 542
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 255.31771850585938,
      "epoch": 0.8143982002249719,
      "grad_norm": 2.0654717881411986,
      "kl": 1.12646484375,
      "learning_rate": 2.009613570788057e-06,
      "loss": 0.1574,
      "reward": 2.770683705806732,
      "reward_std": 0.2916657757014036,
      "rewards/accuracy_reward": 0.8177083730697632,
      "rewards/reasoning_steps_reward": 1.0,
      "rewards/repetition_penalty_reward": -0.03660803521052003,
      "rewards/tag_count_reward": 0.9895833432674408,
      "step": 543
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 298.3854217529297,
      "epoch": 0.8158980127484065,
      "grad_norm": 3.1495657832552553,
      "kl": 2.18017578125,
      "learning_rate": 1.978188122326683e-06,
      "loss": 0.1062,
      "reward": 2.6357369422912598,
      "reward_std": 0.3864106610417366,
      "rewards/accuracy_reward": 0.723958358168602,
      "rewards/reasoning_steps_reward": 0.9843750149011612,
      "rewards/repetition_penalty_reward": -0.053065232932567596,
      "rewards/tag_count_reward": 0.9804687798023224,
      "step": 544
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 277.87501525878906,
      "epoch": 0.817397825271841,
      "grad_norm": 4.472839155741773,
      "kl": 2.248046875,
      "learning_rate": 1.946983330798621e-06,
      "loss": 0.1409,
      "reward": 2.5282241702079773,
      "reward_std": 0.4547024220228195,
      "rewards/accuracy_reward": 0.640625,
      "rewards/reasoning_steps_reward": 0.973958358168602,
      "rewards/repetition_penalty_reward": -0.043390512466430664,
      "rewards/tag_count_reward": 0.9570312798023224,
      "step": 545
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 247.5989646911621,
      "epoch": 0.8188976377952756,
      "grad_norm": 24.974566214450327,
      "kl": 2.99853515625,
      "learning_rate": 1.916000054557783e-06,
      "loss": 0.4573,
      "reward": 2.6690892577171326,
      "reward_std": 0.4272429645061493,
      "rewards/accuracy_reward": 0.7447916865348816,
      "rewards/reasoning_steps_reward": 0.9895833432674408,
      "rewards/repetition_penalty_reward": -0.03794198762625456,
      "rewards/tag_count_reward": 0.9726562798023224,
      "step": 546
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 299.86458587646484,
      "epoch": 0.8203974503187101,
      "grad_norm": 10.956923315304682,
      "kl": 5.841796875,
      "learning_rate": 1.8852391458648323e-06,
      "loss": 0.3112,
      "reward": 2.711550772190094,
      "reward_std": 0.4324747771024704,
      "rewards/accuracy_reward": 0.817708358168602,
      "rewards/reasoning_steps_reward": 0.9843750149011612,
      "rewards/repetition_penalty_reward": -0.04886595252901316,
      "rewards/tag_count_reward": 0.9583333432674408,
      "step": 547
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 264.12500762939453,
      "epoch": 0.8218972628421447,
      "grad_norm": 10.26335940256424,
      "kl": 3.107421875,
      "learning_rate": 1.854701450863744e-06,
      "loss": 0.3093,
      "reward": 2.495303452014923,
      "reward_std": 0.406343936920166,
      "rewards/accuracy_reward": 0.6093750149011612,
      "rewards/reasoning_steps_reward": 0.9791666865348816,
      "rewards/repetition_penalty_reward": -0.0489674867130816,
      "rewards/tag_count_reward": 0.9557291865348816,
      "step": 548
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 268.10938262939453,
      "epoch": 0.8233970753655793,
      "grad_norm": 3.7641473071951963,
      "kl": 3.146484375,
      "learning_rate": 1.8243878095585244e-06,
      "loss": 0.1475,
      "reward": 2.7090840935707092,
      "reward_std": 0.482110857963562,
      "rewards/accuracy_reward": 0.8281250298023224,
      "rewards/reasoning_steps_reward": 0.9791666865348816,
      "rewards/repetition_penalty_reward": -0.04872854123823345,
      "rewards/tag_count_reward": 0.9505208432674408,
      "step": 549
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 293.93750762939453,
      "epoch": 0.8248968878890138,
      "grad_norm": 4.346280398373657,
      "kl": 1.73583984375,
      "learning_rate": 1.7942990557901119e-06,
      "loss": 0.1208,
      "reward": 2.5294079780578613,
      "reward_std": 0.44582589715719223,
      "rewards/accuracy_reward": 0.6197916939854622,
      "rewards/reasoning_steps_reward": 0.9895833432674408,
      "rewards/repetition_penalty_reward": -0.05262336507439613,
      "rewards/tag_count_reward": 0.9726562649011612,
      "step": 550
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 268.5572967529297,
      "epoch": 0.8263967004124484,
      "grad_norm": 12.194807976372957,
      "kl": 0.8740234375,
      "learning_rate": 1.7644360172134323e-06,
      "loss": 0.2462,
      "reward": 2.5935970544815063,
      "reward_std": 0.3754644878208637,
      "rewards/accuracy_reward": 0.6666667014360428,
      "rewards/reasoning_steps_reward": 0.9895833432674408,
      "rewards/repetition_penalty_reward": -0.044423856772482395,
      "rewards/tag_count_reward": 0.981770858168602,
      "step": 551
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 271.2864761352539,
      "epoch": 0.8278965129358831,
      "grad_norm": 3.492663781225749,
      "kl": 0.78857421875,
      "learning_rate": 1.734799515274641e-06,
      "loss": 0.1673,
      "reward": 2.57718825340271,
      "reward_std": 0.3006473407149315,
      "rewards/accuracy_reward": 0.6562500149011612,
      "rewards/reasoning_steps_reward": 0.9913194477558136,
      "rewards/repetition_penalty_reward": -0.0521521158516407,
      "rewards/tag_count_reward": 0.981770858168602,
      "step": 552
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 206.7031364440918,
      "epoch": 0.8293963254593176,
      "grad_norm": 1.0800649313598587,
      "kl": 0.86865234375,
      "learning_rate": 1.7053903651885217e-06,
      "loss": 0.069,
      "reward": 2.665630042552948,
      "reward_std": 0.27732831984758377,
      "rewards/accuracy_reward": 0.7135416865348816,
      "rewards/reasoning_steps_reward": 0.9930555522441864,
      "rewards/repetition_penalty_reward": -0.02924854727461934,
      "rewards/tag_count_reward": 0.9882812649011612,
      "step": 553
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 255.51562881469727,
      "epoch": 0.8308961379827522,
      "grad_norm": 6.658607380817447,
      "kl": 2.01953125,
      "learning_rate": 1.6762093759160614e-06,
      "loss": 0.247,
      "reward": 2.5456653237342834,
      "reward_std": 0.5418491065502167,
      "rewards/accuracy_reward": 0.6718750298023224,
      "rewards/reasoning_steps_reward": 0.9687500149011612,
      "rewards/repetition_penalty_reward": -0.051991009153425694,
      "rewards/tag_count_reward": 0.9570312649011612,
      "step": 554
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 228.35938262939453,
      "epoch": 0.8323959505061868,
      "grad_norm": 5.821343444236744,
      "kl": 2.45849609375,
      "learning_rate": 1.647257350142204e-06,
      "loss": 0.1243,
      "reward": 2.5485109090805054,
      "reward_std": 0.4390419125556946,
      "rewards/accuracy_reward": 0.645833358168602,
      "rewards/reasoning_steps_reward": 0.9791666716337204,
      "rewards/repetition_penalty_reward": -0.04133300017565489,
      "rewards/tag_count_reward": 0.9648437649011612,
      "step": 555
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 226.7291717529297,
      "epoch": 0.8338957630296213,
      "grad_norm": 1.79998365352896,
      "kl": 1.49853515625,
      "learning_rate": 1.618535084253765e-06,
      "loss": 0.1319,
      "reward": 2.61370986700058,
      "reward_std": 0.3187000434845686,
      "rewards/accuracy_reward": 0.6822916716337204,
      "rewards/reasoning_steps_reward": 0.9930555522441864,
      "rewards/repetition_penalty_reward": -0.03689793427474797,
      "rewards/tag_count_reward": 0.9752604365348816,
      "step": 556
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 230.59375762939453,
      "epoch": 0.8353955755530559,
      "grad_norm": 16.824204922339664,
      "kl": 3.234375,
      "learning_rate": 1.5900433683175277e-06,
      "loss": 0.5503,
      "reward": 2.6835938692092896,
      "reward_std": 0.4415631741285324,
      "rewards/accuracy_reward": 0.7604166865348816,
      "rewards/reasoning_steps_reward": 0.9930555522441864,
      "rewards/repetition_penalty_reward": -0.04253482352942228,
      "rewards/tag_count_reward": 0.9726562649011612,
      "step": 557
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 225.4375114440918,
      "epoch": 0.8368953880764904,
      "grad_norm": 8.685664685007387,
      "kl": 2.427734375,
      "learning_rate": 1.5617829860585087e-06,
      "loss": 0.1994,
      "reward": 2.531205892562866,
      "reward_std": 0.43269092589616776,
      "rewards/accuracy_reward": 0.6093750298023224,
      "rewards/reasoning_steps_reward": 0.986111119389534,
      "rewards/repetition_penalty_reward": -0.03172821528278291,
      "rewards/tag_count_reward": 0.9674479365348816,
      "step": 558
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 208.06250762939453,
      "epoch": 0.838395200599925,
      "grad_norm": 11.917922673045826,
      "kl": 2.44091796875,
      "learning_rate": 1.533754714838408e-06,
      "loss": 0.1954,
      "reward": 2.4975586533546448,
      "reward_std": 0.44412345439195633,
      "rewards/accuracy_reward": 0.5937500149011612,
      "rewards/reasoning_steps_reward": 0.9635416865348816,
      "rewards/repetition_penalty_reward": -0.02848310861736536,
      "rewards/tag_count_reward": 0.9687500298023224,
      "step": 559
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 259.4427185058594,
      "epoch": 0.8398950131233596,
      "grad_norm": 4.400151882822108,
      "kl": 2.544921875,
      "learning_rate": 1.5059593256342142e-06,
      "loss": 0.0563,
      "reward": 2.6064106225967407,
      "reward_std": 0.4480983540415764,
      "rewards/accuracy_reward": 0.7187500149011612,
      "rewards/reasoning_steps_reward": 0.9722222238779068,
      "rewards/repetition_penalty_reward": -0.05331165995448828,
      "rewards/tag_count_reward": 0.9687500298023224,
      "step": 560
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 234.78646850585938,
      "epoch": 0.8413948256467941,
      "grad_norm": 4.108382543830119,
      "kl": 3.171875,
      "learning_rate": 1.4783975830170028e-06,
      "loss": 0.6147,
      "reward": 2.4788121581077576,
      "reward_std": 0.5859697312116623,
      "rewards/accuracy_reward": 0.614583358168602,
      "rewards/reasoning_steps_reward": 0.967013880610466,
      "rewards/repetition_penalty_reward": -0.046795559115707874,
      "rewards/tag_count_reward": 0.9440104514360428,
      "step": 561
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 219.51042938232422,
      "epoch": 0.8428946381702287,
      "grad_norm": 3.9632943549968522,
      "kl": 2.30029296875,
      "learning_rate": 1.4510702451309055e-06,
      "loss": 0.2233,
      "reward": 2.5965715646743774,
      "reward_std": 0.48612427711486816,
      "rewards/accuracy_reward": 0.7239583730697632,
      "rewards/reasoning_steps_reward": 0.9600694924592972,
      "rewards/repetition_penalty_reward": -0.04578963201493025,
      "rewards/tag_count_reward": 0.958333358168602,
      "step": 562
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 234.33855438232422,
      "epoch": 0.8443944506936633,
      "grad_norm": 12.689273098315706,
      "kl": 2.89453125,
      "learning_rate": 1.4239780636722555e-06,
      "loss": 0.5173,
      "reward": 2.568632483482361,
      "reward_std": 0.4845578037202358,
      "rewards/accuracy_reward": 0.6875000074505806,
      "rewards/reasoning_steps_reward": 0.9756944924592972,
      "rewards/repetition_penalty_reward": -0.05029124254360795,
      "rewards/tag_count_reward": 0.9557292014360428,
      "step": 563
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 255.00521087646484,
      "epoch": 0.8458942632170978,
      "grad_norm": 22.887883651104566,
      "kl": 2.71875,
      "learning_rate": 1.39712178386891e-06,
      "loss": 0.7815,
      "reward": 2.606330990791321,
      "reward_std": 0.6394063234329224,
      "rewards/accuracy_reward": 0.7500000298023224,
      "rewards/reasoning_steps_reward": 0.9618056118488312,
      "rewards/repetition_penalty_reward": -0.05859959963709116,
      "rewards/tag_count_reward": 0.9531250298023224,
      "step": 564
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 291.60939025878906,
      "epoch": 0.8473940757405324,
      "grad_norm": 26.21108108487782,
      "kl": 7.546875,
      "learning_rate": 1.3705021444597521e-06,
      "loss": 1.1027,
      "reward": 2.4728458523750305,
      "reward_std": 0.6928077340126038,
      "rewards/accuracy_reward": 0.6406250149011612,
      "rewards/reasoning_steps_reward": 0.96180559694767,
      "rewards/repetition_penalty_reward": -0.07489718683063984,
      "rewards/tag_count_reward": 0.9453125149011612,
      "step": 565
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 201.21875381469727,
      "epoch": 0.8488938882639671,
      "grad_norm": 3.5925945801436803,
      "kl": 2.39453125,
      "learning_rate": 1.344119877674368e-06,
      "loss": 0.301,
      "reward": 2.4886980652809143,
      "reward_std": 0.45957837253808975,
      "rewards/accuracy_reward": 0.5885416865348816,
      "rewards/reasoning_steps_reward": 0.9687500149011612,
      "rewards/repetition_penalty_reward": -0.036041736137121916,
      "rewards/tag_count_reward": 0.9674479216337204,
      "step": 566
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 256.2395896911621,
      "epoch": 0.8503937007874016,
      "grad_norm": 20.274471532130267,
      "kl": 5.91796875,
      "learning_rate": 1.3179757092129087e-06,
      "loss": 1.0635,
      "reward": 2.3603609800338745,
      "reward_std": 0.6862609535455704,
      "rewards/accuracy_reward": 0.5625000260770321,
      "rewards/reasoning_steps_reward": 0.9513889104127884,
      "rewards/repetition_penalty_reward": -0.06238204799592495,
      "rewards/tag_count_reward": 0.9088541865348816,
      "step": 567
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 226.59375762939453,
      "epoch": 0.8518935133108362,
      "grad_norm": 9.701437197347186,
      "kl": 8.5390625,
      "learning_rate": 1.292070358226124e-06,
      "loss": 0.3018,
      "reward": 2.3854750990867615,
      "reward_std": 0.6592359095811844,
      "rewards/accuracy_reward": 0.5416666939854622,
      "rewards/reasoning_steps_reward": 0.9565972536802292,
      "rewards/repetition_penalty_reward": -0.05028895568102598,
      "rewards/tag_count_reward": 0.9375000298023224,
      "step": 568
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 210.29687881469727,
      "epoch": 0.8533933258342707,
      "grad_norm": 10.655977657198079,
      "kl": 3.45703125,
      "learning_rate": 1.2664045372955858e-06,
      "loss": 0.6047,
      "reward": 2.418867766857147,
      "reward_std": 0.6335871517658234,
      "rewards/accuracy_reward": 0.5729166865348816,
      "rewards/reasoning_steps_reward": 0.9548611491918564,
      "rewards/repetition_penalty_reward": -0.045107895508408546,
      "rewards/tag_count_reward": 0.9361979365348816,
      "step": 569
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 277.40625762939453,
      "epoch": 0.8548931383577053,
      "grad_norm": 11.027587159101916,
      "kl": 6.21484375,
      "learning_rate": 1.2409789524140813e-06,
      "loss": 0.854,
      "reward": 2.5739906430244446,
      "reward_std": 0.7429305166006088,
      "rewards/accuracy_reward": 0.770833358168602,
      "rewards/reasoning_steps_reward": 0.935763955116272,
      "rewards/repetition_penalty_reward": -0.06750249583274126,
      "rewards/tag_count_reward": 0.934895858168602,
      "step": 570
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 266.0937614440918,
      "epoch": 0.8563929508811399,
      "grad_norm": 14.64342890452312,
      "kl": 9.6875,
      "learning_rate": 1.2157943029661977e-06,
      "loss": 0.9433,
      "reward": 2.343918889760971,
      "reward_std": 0.8442835658788681,
      "rewards/accuracy_reward": 0.6510416716337204,
      "rewards/reasoning_steps_reward": 0.8923611491918564,
      "rewards/repetition_penalty_reward": -0.0770882060751319,
      "rewards/tag_count_reward": 0.8776041865348816,
      "step": 571
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 190.70312881469727,
      "epoch": 0.8578927634045744,
      "grad_norm": 4.726128645967551,
      "kl": 1.662109375,
      "learning_rate": 1.1908512817090833e-06,
      "loss": 0.2499,
      "reward": 2.4979942440986633,
      "reward_std": 0.40954509750008583,
      "rewards/accuracy_reward": 0.5937500223517418,
      "rewards/reasoning_steps_reward": 0.9826389253139496,
      "rewards/repetition_penalty_reward": -0.032821862027049065,
      "rewards/tag_count_reward": 0.954427108168602,
      "step": 572
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 229.52605056762695,
      "epoch": 0.859392575928009,
      "grad_norm": 5.56550512002148,
      "kl": 2.75390625,
      "learning_rate": 1.1661505747533897e-06,
      "loss": 0.3795,
      "reward": 2.5485902428627014,
      "reward_std": 0.7243528515100479,
      "rewards/accuracy_reward": 0.7343750149011612,
      "rewards/reasoning_steps_reward": 0.9288194626569748,
      "rewards/repetition_penalty_reward": -0.048198044300079346,
      "rewards/tag_count_reward": 0.9335937649011612,
      "step": 573
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 243.60938262939453,
      "epoch": 0.8608923884514436,
      "grad_norm": 7.741451341316265,
      "kl": 3.44140625,
      "learning_rate": 1.1416928615444013e-06,
      "loss": 0.5439,
      "reward": 2.4548073410987854,
      "reward_std": 0.7477606385946274,
      "rewards/accuracy_reward": 0.6406250298023224,
      "rewards/reasoning_steps_reward": 0.9531250596046448,
      "rewards/repetition_penalty_reward": -0.05040103476494551,
      "rewards/tag_count_reward": 0.911458358168602,
      "step": 574
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 239.36459350585938,
      "epoch": 0.8623922009748781,
      "grad_norm": 12.69637648259022,
      "kl": 5.462890625,
      "learning_rate": 1.1174788148433423e-06,
      "loss": 0.7872,
      "reward": 2.5563725233078003,
      "reward_std": 0.7221258133649826,
      "rewards/accuracy_reward": 0.7760416716337204,
      "rewards/reasoning_steps_reward": 0.9184028208255768,
      "rewards/repetition_penalty_reward": -0.0625512283295393,
      "rewards/tag_count_reward": 0.9244792014360428,
      "step": 575
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 232.67708587646484,
      "epoch": 0.8638920134983127,
      "grad_norm": 6.651240220699383,
      "kl": 2.841796875,
      "learning_rate": 1.0935091007088761e-06,
      "loss": 0.5463,
      "reward": 2.4423410296440125,
      "reward_std": 0.6437135934829712,
      "rewards/accuracy_reward": 0.619791679084301,
      "rewards/reasoning_steps_reward": 0.9409722238779068,
      "rewards/repetition_penalty_reward": -0.0559229115024209,
      "rewards/tag_count_reward": 0.9375000298023224,
      "step": 576
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 215.6822967529297,
      "epoch": 0.8653918260217472,
      "grad_norm": 9.069503918590028,
      "kl": 3.73046875,
      "learning_rate": 1.069784378478781e-06,
      "loss": 0.5511,
      "reward": 2.417185962200165,
      "reward_std": 0.6344876810908318,
      "rewards/accuracy_reward": 0.6093750223517418,
      "rewards/reasoning_steps_reward": 0.9340277761220932,
      "rewards/repetition_penalty_reward": -0.04809193778783083,
      "rewards/tag_count_reward": 0.9218750149011612,
      "step": 577
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 210.9166717529297,
      "epoch": 0.8668916385451818,
      "grad_norm": 3.9657531302416285,
      "kl": 2.365234375,
      "learning_rate": 1.046305300751811e-06,
      "loss": 0.2354,
      "reward": 2.5444520115852356,
      "reward_std": 0.678747646510601,
      "rewards/accuracy_reward": 0.7135416865348816,
      "rewards/reasoning_steps_reward": 0.9531250149011612,
      "rewards/repetition_penalty_reward": -0.0466939271427691,
      "rewards/tag_count_reward": 0.9244791716337204,
      "step": 578
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 248.43750381469727,
      "epoch": 0.8683914510686164,
      "grad_norm": 13.803690175273248,
      "kl": 3.46484375,
      "learning_rate": 1.0230725133697495e-06,
      "loss": 0.7416,
      "reward": 2.505746603012085,
      "reward_std": 0.6760745644569397,
      "rewards/accuracy_reward": 0.692708358168602,
      "rewards/reasoning_steps_reward": 0.947916716337204,
      "rewards/repetition_penalty_reward": -0.05935762915760279,
      "rewards/tag_count_reward": 0.9244792014360428,
      "step": 579
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 222.7552146911621,
      "epoch": 0.869891263592051,
      "grad_norm": 6.038364170548584,
      "kl": 1.857421875,
      "learning_rate": 1.0000866553996436e-06,
      "loss": 0.3797,
      "reward": 2.563169300556183,
      "reward_std": 0.49185874313116074,
      "rewards/accuracy_reward": 0.6822916865348816,
      "rewards/reasoning_steps_reward": 0.958333358168602,
      "rewards/repetition_penalty_reward": -0.0383933181874454,
      "rewards/tag_count_reward": 0.9609375298023224,
      "step": 580
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 206.24480056762695,
      "epoch": 0.8713910761154856,
      "grad_norm": 2.7457030988838373,
      "kl": 2.619140625,
      "learning_rate": 9.773483591162203e-07,
      "loss": 0.3962,
      "reward": 2.5448489785194397,
      "reward_std": 0.5937497019767761,
      "rewards/accuracy_reward": 0.6875000149011612,
      "rewards/reasoning_steps_reward": 0.9635417014360428,
      "rewards/repetition_penalty_reward": -0.04499485623091459,
      "rewards/tag_count_reward": 0.938802108168602,
      "step": 581
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 204.1354217529297,
      "epoch": 0.8728908886389202,
      "grad_norm": 3.9924980474405847,
      "kl": 3.34765625,
      "learning_rate": 9.548582499845015e-07,
      "loss": 0.0155,
      "reward": 2.46910959482193,
      "reward_std": 0.7316218465566635,
      "rewards/accuracy_reward": 0.6718750298023224,
      "rewards/reasoning_steps_reward": 0.9114583730697632,
      "rewards/repetition_penalty_reward": -0.04260919149965048,
      "rewards/tag_count_reward": 0.9283854365348816,
      "step": 582
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 185.3177146911621,
      "epoch": 0.8743907011623547,
      "grad_norm": 3.3857582260464882,
      "kl": 1.8125,
      "learning_rate": 9.326169466425916e-07,
      "loss": 0.2775,
      "reward": 2.697448194026947,
      "reward_std": 0.5078239142894745,
      "rewards/accuracy_reward": 0.7760416865348816,
      "rewards/reasoning_steps_reward": 0.9843750149011612,
      "rewards/repetition_penalty_reward": -0.03432267066091299,
      "rewards/tag_count_reward": 0.9713541865348816,
      "step": 583
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 206.1875,
      "epoch": 0.8758905136857893,
      "grad_norm": 11.742745507237299,
      "kl": 2.9140625,
      "learning_rate": 9.106250608846679e-07,
      "loss": 0.5072,
      "reward": 2.6075586080551147,
      "reward_std": 0.620074912905693,
      "rewards/accuracy_reward": 0.7343750149011612,
      "rewards/reasoning_steps_reward": 0.9600694924592972,
      "rewards/repetition_penalty_reward": -0.036104677245020866,
      "rewards/tag_count_reward": 0.9492187649011612,
      "step": 584
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 217.8541717529297,
      "epoch": 0.8773903262092239,
      "grad_norm": 2.2041577561531684,
      "kl": 2.1953125,
      "learning_rate": 8.888831976441481e-07,
      "loss": 0.4111,
      "reward": 2.6042516231536865,
      "reward_std": 0.47759709507226944,
      "rewards/accuracy_reward": 0.6822916865348816,
      "rewards/reasoning_steps_reward": 0.9809028059244156,
      "rewards/repetition_penalty_reward": -0.03159915190190077,
      "rewards/tag_count_reward": 0.9726562649011612,
      "step": 585
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 205.8229217529297,
      "epoch": 0.8788901387326584,
      "grad_norm": 4.274612409465976,
      "kl": 2.3701171875,
      "learning_rate": 8.673919549770483e-07,
      "loss": 0.3398,
      "reward": 2.611320972442627,
      "reward_std": 0.5121708884835243,
      "rewards/accuracy_reward": 0.7291666865348816,
      "rewards/reasoning_steps_reward": 0.960069477558136,
      "rewards/repetition_penalty_reward": -0.044061001390218735,
      "rewards/tag_count_reward": 0.9661458432674408,
      "step": 586
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 202.00521087646484,
      "epoch": 0.880389951256093,
      "grad_norm": 3.736601399113756,
      "kl": 2.330078125,
      "learning_rate": 8.461519240455362e-07,
      "loss": 0.3635,
      "reward": 2.6235088109970093,
      "reward_std": 0.5086311176419258,
      "rewards/accuracy_reward": 0.7187500149011612,
      "rewards/reasoning_steps_reward": 0.986111119389534,
      "rewards/repetition_penalty_reward": -0.037081570364534855,
      "rewards/tag_count_reward": 0.9557291716337204,
      "step": 587
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 209.96355056762695,
      "epoch": 0.8818897637795275,
      "grad_norm": 17.45784244105432,
      "kl": 20.078125,
      "learning_rate": 8.251636891016702e-07,
      "loss": 0.5062,
      "reward": 2.5059803128242493,
      "reward_std": 0.631468877196312,
      "rewards/accuracy_reward": 0.6354166865348816,
      "rewards/reasoning_steps_reward": 0.9618055522441864,
      "rewards/repetition_penalty_reward": -0.037856731098145247,
      "rewards/tag_count_reward": 0.9466146230697632,
      "step": 588
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 219.26042556762695,
      "epoch": 0.8833895763029621,
      "grad_norm": 2.4924144055714925,
      "kl": 2.31640625,
      "learning_rate": 8.044278274713246e-07,
      "loss": 0.6179,
      "reward": 2.7582138180732727,
      "reward_std": 0.4812832549214363,
      "rewards/accuracy_reward": 0.8697916865348816,
      "rewards/reasoning_steps_reward": 0.9722222685813904,
      "rewards/repetition_penalty_reward": -0.048643797636032104,
      "rewards/tag_count_reward": 0.9648437649011612,
      "step": 589
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 213.8072967529297,
      "epoch": 0.8848893888263967,
      "grad_norm": 4.087042616486125,
      "kl": 2.037109375,
      "learning_rate": 7.839449095383111e-07,
      "loss": 0.4912,
      "reward": 2.69020676612854,
      "reward_std": 0.5107688158750534,
      "rewards/accuracy_reward": 0.786458358168602,
      "rewards/reasoning_steps_reward": 0.9791666865348816,
      "rewards/repetition_penalty_reward": -0.04026202671229839,
      "rewards/tag_count_reward": 0.9648437798023224,
      "step": 590
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 220.7447967529297,
      "epoch": 0.8863892013498312,
      "grad_norm": 14.915815383649845,
      "kl": 2.236328125,
      "learning_rate": 7.637154987286888e-07,
      "loss": 0.4128,
      "reward": 2.6318877935409546,
      "reward_std": 0.41991502046585083,
      "rewards/accuracy_reward": 0.7343750298023224,
      "rewards/reasoning_steps_reward": 0.9826389104127884,
      "rewards/repetition_penalty_reward": -0.04476172663271427,
      "rewards/tag_count_reward": 0.9596354365348816,
      "step": 591
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 212.46875762939453,
      "epoch": 0.8878890138732658,
      "grad_norm": 6.050727563412662,
      "kl": 2.1787109375,
      "learning_rate": 7.437401514952646e-07,
      "loss": 0.3731,
      "reward": 2.4536707997322083,
      "reward_std": 0.4574524015188217,
      "rewards/accuracy_reward": 0.5729166716337204,
      "rewards/reasoning_steps_reward": 0.9704861342906952,
      "rewards/repetition_penalty_reward": -0.04676327481865883,
      "rewards/tag_count_reward": 0.9570312798023224,
      "step": 592
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 222.2552146911621,
      "epoch": 0.8893888263967004,
      "grad_norm": 9.16124525709699,
      "kl": 4.005859375,
      "learning_rate": 7.240194173022941e-07,
      "loss": 0.5731,
      "reward": 2.6355690360069275,
      "reward_std": 0.5524905323982239,
      "rewards/accuracy_reward": 0.7812500149011612,
      "rewards/reasoning_steps_reward": 0.963541716337204,
      "rewards/repetition_penalty_reward": -0.05713945999741554,
      "rewards/tag_count_reward": 0.9479166865348816,
      "step": 593
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 231.3489646911621,
      "epoch": 0.890888638920135,
      "grad_norm": 22.380535968926747,
      "kl": 3.015625,
      "learning_rate": 7.045538386103579e-07,
      "loss": 0.4854,
      "reward": 2.4118546843528748,
      "reward_std": 0.6012802645564079,
      "rewards/accuracy_reward": 0.5677083507180214,
      "rewards/reasoning_steps_reward": 0.9600694477558136,
      "rewards/repetition_penalty_reward": -0.057329487055540085,
      "rewards/tag_count_reward": 0.9414062798023224,
      "step": 594
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 268.4322967529297,
      "epoch": 0.8923884514435696,
      "grad_norm": 3.774203451456834,
      "kl": 3.220703125,
      "learning_rate": 6.853439508614412e-07,
      "loss": 0.6382,
      "reward": 2.497507393360138,
      "reward_std": 0.5804141908884048,
      "rewards/accuracy_reward": 0.6510416865348816,
      "rewards/reasoning_steps_reward": 0.9739583283662796,
      "rewards/repetition_penalty_reward": -0.06759683508425951,
      "rewards/tag_count_reward": 0.9401041865348816,
      "step": 595
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 195.17709350585938,
      "epoch": 0.8938882639670042,
      "grad_norm": 4.560347288059126,
      "kl": 2.53125,
      "learning_rate": 6.663902824642132e-07,
      "loss": 0.2865,
      "reward": 2.6934563517570496,
      "reward_std": 0.5186980329453945,
      "rewards/accuracy_reward": 0.8281250149011612,
      "rewards/reasoning_steps_reward": 0.9496527761220932,
      "rewards/repetition_penalty_reward": -0.03614447545260191,
      "rewards/tag_count_reward": 0.9518229365348816,
      "step": 596
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 251.75,
      "epoch": 0.8953880764904387,
      "grad_norm": 20.778553368266753,
      "kl": 2.6484375,
      "learning_rate": 6.47693354779484e-07,
      "loss": 0.6244,
      "reward": 2.6013853549957275,
      "reward_std": 0.6506573259830475,
      "rewards/accuracy_reward": 0.7552083507180214,
      "rewards/reasoning_steps_reward": 0.9635416716337204,
      "rewards/repetition_penalty_reward": -0.0626771878451109,
      "rewards/tag_count_reward": 0.9453125,
      "step": 597
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 249.87500762939453,
      "epoch": 0.8968878890138733,
      "grad_norm": 9.731098045072335,
      "kl": 2.041015625,
      "learning_rate": 6.29253682105866e-07,
      "loss": 0.4267,
      "reward": 2.6623805165290833,
      "reward_std": 0.5037910491228104,
      "rewards/accuracy_reward": 0.7916667014360428,
      "rewards/reasoning_steps_reward": 0.9809028059244156,
      "rewards/repetition_penalty_reward": -0.05940787214785814,
      "rewards/tag_count_reward": 0.9492187649011612,
      "step": 598
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 239.6197967529297,
      "epoch": 0.8983877015373078,
      "grad_norm": 8.23836216740641,
      "kl": 2.60546875,
      "learning_rate": 6.110717716656289e-07,
      "loss": 0.4356,
      "reward": 2.4078712463378906,
      "reward_std": 0.6441041380167007,
      "rewards/accuracy_reward": 0.5677083432674408,
      "rewards/reasoning_steps_reward": 0.9496528059244156,
      "rewards/repetition_penalty_reward": -0.06001073541119695,
      "rewards/tag_count_reward": 0.9505208432674408,
      "step": 599
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 233.6458396911621,
      "epoch": 0.8998875140607424,
      "grad_norm": 5.950701052370844,
      "kl": 3.33203125,
      "learning_rate": 5.931481235907466e-07,
      "loss": 0.5106,
      "reward": 2.4382471442222595,
      "reward_std": 0.7465489208698273,
      "rewards/accuracy_reward": 0.6406250149011612,
      "rewards/reasoning_steps_reward": 0.927083358168602,
      "rewards/repetition_penalty_reward": -0.05394035950303078,
      "rewards/tag_count_reward": 0.9244791865348816,
      "step": 600
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 232.4635467529297,
      "epoch": 0.901387326584177,
      "grad_norm": 8.835485931240282,
      "kl": 3.53515625,
      "learning_rate": 5.754832309091362e-07,
      "loss": 0.6214,
      "reward": 2.6882564425468445,
      "reward_std": 0.6361054480075836,
      "rewards/accuracy_reward": 0.833333358168602,
      "rewards/reasoning_steps_reward": 0.9670139253139496,
      "rewards/repetition_penalty_reward": -0.058705421164631844,
      "rewards/tag_count_reward": 0.946614608168602,
      "step": 601
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 213.64584350585938,
      "epoch": 0.9028871391076115,
      "grad_norm": 4.936233068343242,
      "kl": 2.7421875,
      "learning_rate": 5.580775795311033e-07,
      "loss": 0.3129,
      "reward": 2.5619428157806396,
      "reward_std": 0.6096071302890778,
      "rewards/accuracy_reward": 0.713541679084301,
      "rewards/reasoning_steps_reward": 0.9513889402151108,
      "rewards/repetition_penalty_reward": -0.04439397854730487,
      "rewards/tag_count_reward": 0.9414062798023224,
      "step": 602
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 226.42188262939453,
      "epoch": 0.9043869516310461,
      "grad_norm": 4.234086119916309,
      "kl": 3.302734375,
      "learning_rate": 5.409316482359694e-07,
      "loss": 0.4198,
      "reward": 2.5895190834999084,
      "reward_std": 0.5498283728957176,
      "rewards/accuracy_reward": 0.7083333432674408,
      "rewards/reasoning_steps_reward": 0.973958358168602,
      "rewards/repetition_penalty_reward": -0.0471997894346714,
      "rewards/tag_count_reward": 0.954427108168602,
      "step": 603
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 213.32291793823242,
      "epoch": 0.9058867641544807,
      "grad_norm": 11.124045036921478,
      "kl": 3.056640625,
      "learning_rate": 5.240459086589056e-07,
      "loss": 0.5874,
      "reward": 2.640231668949127,
      "reward_std": 0.507370337843895,
      "rewards/accuracy_reward": 0.7656250298023224,
      "rewards/reasoning_steps_reward": 0.9722222089767456,
      "rewards/repetition_penalty_reward": -0.04683440877124667,
      "rewards/tag_count_reward": 0.9492187649011612,
      "step": 604
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 252.64064025878906,
      "epoch": 0.9073865766779152,
      "grad_norm": 8.760847362194967,
      "kl": 3.0859375,
      "learning_rate": 5.074208252779589e-07,
      "loss": 0.6574,
      "reward": 2.524926483631134,
      "reward_std": 0.6083704084157944,
      "rewards/accuracy_reward": 0.630208358168602,
      "rewards/reasoning_steps_reward": 0.9809028059244156,
      "rewards/repetition_penalty_reward": -0.048424381762742996,
      "rewards/tag_count_reward": 0.962239608168602,
      "step": 605
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 255.46355056762695,
      "epoch": 0.9088863892013498,
      "grad_norm": 10.952534285750453,
      "kl": 3.94921875,
      "learning_rate": 4.910568554012751e-07,
      "loss": 0.7252,
      "reward": 2.4818572998046875,
      "reward_std": 0.5148201733827591,
      "rewards/accuracy_reward": 0.5937500149011612,
      "rewards/reasoning_steps_reward": 0.9878472238779068,
      "rewards/repetition_penalty_reward": -0.06067758519202471,
      "rewards/tag_count_reward": 0.9609375298023224,
      "step": 606
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 201.06771087646484,
      "epoch": 0.9103862017247843,
      "grad_norm": 9.032595902462477,
      "kl": 3.17578125,
      "learning_rate": 4.749544491545199e-07,
      "loss": 0.5553,
      "reward": 2.5263285636901855,
      "reward_std": 0.5500727593898773,
      "rewards/accuracy_reward": 0.6562500223517418,
      "rewards/reasoning_steps_reward": 0.9687500149011612,
      "rewards/repetition_penalty_reward": -0.046588233672082424,
      "rewards/tag_count_reward": 0.9479166865348816,
      "step": 607
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 234.6510467529297,
      "epoch": 0.911886014248219,
      "grad_norm": 3.4371655420891853,
      "kl": 2.94140625,
      "learning_rate": 4.591140494684965e-07,
      "loss": 0.5164,
      "reward": 2.340811610221863,
      "reward_std": 0.4512625113129616,
      "rewards/accuracy_reward": 0.4635416865348816,
      "rewards/reasoning_steps_reward": 0.9774305671453476,
      "rewards/repetition_penalty_reward": -0.05198352737352252,
      "rewards/tag_count_reward": 0.9518229365348816,
      "step": 608
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 238.08333587646484,
      "epoch": 0.9133858267716536,
      "grad_norm": 5.5444529742596265,
      "kl": 2.482421875,
      "learning_rate": 4.435360920669618e-07,
      "loss": 0.4373,
      "reward": 2.7700709104537964,
      "reward_std": 0.48688896745443344,
      "rewards/accuracy_reward": 0.8750000149011612,
      "rewards/reasoning_steps_reward": 0.9809028059244156,
      "rewards/repetition_penalty_reward": -0.0454674381762743,
      "rewards/tag_count_reward": 0.9596354216337204,
      "step": 609
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 215.10937881469727,
      "epoch": 0.9148856392950881,
      "grad_norm": 2.692978335091646,
      "kl": 3.86328125,
      "learning_rate": 4.282210054546454e-07,
      "loss": 0.7832,
      "reward": 2.5629987120628357,
      "reward_std": 0.5417983531951904,
      "rewards/accuracy_reward": 0.7031250223517418,
      "rewards/reasoning_steps_reward": 0.9652778059244156,
      "rewards/repetition_penalty_reward": -0.053320798091590405,
      "rewards/tag_count_reward": 0.9479166865348816,
      "step": 610
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 201.1666717529297,
      "epoch": 0.9163854518185227,
      "grad_norm": 6.972427384378082,
      "kl": 2.6015625,
      "learning_rate": 4.1316921090545305e-07,
      "loss": 0.1151,
      "reward": 2.58855801820755,
      "reward_std": 0.4848596230149269,
      "rewards/accuracy_reward": 0.6822916865348816,
      "rewards/reasoning_steps_reward": 0.973958358168602,
      "rewards/repetition_penalty_reward": -0.03513997979462147,
      "rewards/tag_count_reward": 0.9674479365348816,
      "step": 611
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 235.22396850585938,
      "epoch": 0.9178852643419573,
      "grad_norm": 4.93152288537315,
      "kl": 3.115234375,
      "learning_rate": 3.9838112245088934e-07,
      "loss": 0.5991,
      "reward": 2.601243317127228,
      "reward_std": 0.5846336483955383,
      "rewards/accuracy_reward": 0.7395833730697632,
      "rewards/reasoning_steps_reward": 0.9635416716337204,
      "rewards/repetition_penalty_reward": -0.05240248655900359,
      "rewards/tag_count_reward": 0.950520858168602,
      "step": 612
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 262.22396087646484,
      "epoch": 0.9193850768653918,
      "grad_norm": 12.62005524980775,
      "kl": 5.4765625,
      "learning_rate": 3.8385714686866137e-07,
      "loss": 1.1448,
      "reward": 2.5160588026046753,
      "reward_std": 0.7233484238386154,
      "rewards/accuracy_reward": 0.7031250149011612,
      "rewards/reasoning_steps_reward": 0.9531249850988388,
      "rewards/repetition_penalty_reward": -0.06987872999161482,
      "rewards/tag_count_reward": 0.9296875298023224,
      "step": 613
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 235.5989646911621,
      "epoch": 0.9208848893888264,
      "grad_norm": 11.546226463954667,
      "kl": 4.08984375,
      "learning_rate": 3.695976836714932e-07,
      "loss": 0.6332,
      "reward": 2.414819598197937,
      "reward_std": 0.6042335405945778,
      "rewards/accuracy_reward": 0.5677083507180214,
      "rewards/reasoning_steps_reward": 0.9652778208255768,
      "rewards/repetition_penalty_reward": -0.05306240823119879,
      "rewards/tag_count_reward": 0.9348958730697632,
      "step": 614
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 227.06250381469727,
      "epoch": 0.922384701912261,
      "grad_norm": 8.789826083743023,
      "kl": 2.94140625,
      "learning_rate": 3.556031250961356e-07,
      "loss": 0.5855,
      "reward": 2.4650281071662903,
      "reward_std": 0.5545367747545242,
      "rewards/accuracy_reward": 0.6041666865348816,
      "rewards/reasoning_steps_reward": 0.9670139104127884,
      "rewards/repetition_penalty_reward": -0.05146504007279873,
      "rewards/tag_count_reward": 0.9453125,
      "step": 615
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 235.88542556762695,
      "epoch": 0.9238845144356955,
      "grad_norm": 5.27373650876101,
      "kl": 2.8310546875,
      "learning_rate": 3.4187385609257275e-07,
      "loss": 0.4564,
      "reward": 2.5718677043914795,
      "reward_std": 0.5033154115080833,
      "rewards/accuracy_reward": 0.7031250298023224,
      "rewards/reasoning_steps_reward": 0.9791667014360428,
      "rewards/repetition_penalty_reward": -0.06224700016900897,
      "rewards/tag_count_reward": 0.9518229365348816,
      "step": 616
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 254.36458587646484,
      "epoch": 0.9253843269591301,
      "grad_norm": 7.526278127901626,
      "kl": 1.7294921875,
      "learning_rate": 3.284102543134426e-07,
      "loss": 0.3597,
      "reward": 2.546690046787262,
      "reward_std": 0.5290744379162788,
      "rewards/accuracy_reward": 0.6666666865348816,
      "rewards/reasoning_steps_reward": 0.972222238779068,
      "rewards/repetition_penalty_reward": -0.055740559473633766,
      "rewards/tag_count_reward": 0.9635416716337204,
      "step": 617
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 206.28125762939453,
      "epoch": 0.9268841394825647,
      "grad_norm": 7.942979632041828,
      "kl": 5.625,
      "learning_rate": 3.152126901036401e-07,
      "loss": 0.2052,
      "reward": 2.519493043422699,
      "reward_std": 0.4696878641843796,
      "rewards/accuracy_reward": 0.6093750149011612,
      "rewards/reasoning_steps_reward": 0.9791666865348816,
      "rewards/repetition_penalty_reward": -0.035194564145058393,
      "rewards/tag_count_reward": 0.966145858168602,
      "step": 618
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 228.27084350585938,
      "epoch": 0.9283839520059992,
      "grad_norm": 7.258620893937008,
      "kl": 2.21875,
      "learning_rate": 3.0228152649013133e-07,
      "loss": 0.5912,
      "reward": 2.6479897499084473,
      "reward_std": 0.4475140795111656,
      "rewards/accuracy_reward": 0.7500000149011612,
      "rewards/reasoning_steps_reward": 0.9930555671453476,
      "rewards/repetition_penalty_reward": -0.05209715012460947,
      "rewards/tag_count_reward": 0.9570312649011612,
      "step": 619
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 205.40625381469727,
      "epoch": 0.9298837645294338,
      "grad_norm": 4.763052053787403,
      "kl": 2.03466796875,
      "learning_rate": 2.896171191719743e-07,
      "loss": 0.2601,
      "reward": 2.6514768600463867,
      "reward_std": 0.5323489122092724,
      "rewards/accuracy_reward": 0.7552083432674408,
      "rewards/reasoning_steps_reward": 0.9739583432674408,
      "rewards/repetition_penalty_reward": -0.03602320980280638,
      "rewards/tag_count_reward": 0.9583333432674408,
      "step": 620
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 246.7291717529297,
      "epoch": 0.9313835770528683,
      "grad_norm": 4.657452344039985,
      "kl": 3.841796875,
      "learning_rate": 2.772198165105267e-07,
      "loss": 0.9583,
      "reward": 2.4812204241752625,
      "reward_std": 0.6263424828648567,
      "rewards/accuracy_reward": 0.6197916716337204,
      "rewards/reasoning_steps_reward": 0.9722222238779068,
      "rewards/repetition_penalty_reward": -0.06261651404201984,
      "rewards/tag_count_reward": 0.9518229365348816,
      "step": 621
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 234.05730056762695,
      "epoch": 0.932883389576303,
      "grad_norm": 7.137978430111716,
      "kl": 3.41796875,
      "learning_rate": 2.6508995951986526e-07,
      "loss": 0.6458,
      "reward": 2.60908043384552,
      "reward_std": 0.6289382129907608,
      "rewards/accuracy_reward": 0.7656250298023224,
      "rewards/reasoning_steps_reward": 0.9565972685813904,
      "rewards/repetition_penalty_reward": -0.06236053630709648,
      "rewards/tag_count_reward": 0.9492187798023224,
      "step": 622
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 203.58855056762695,
      "epoch": 0.9343832020997376,
      "grad_norm": 8.613599095668501,
      "kl": 2.337890625,
      "learning_rate": 2.532278818574108e-07,
      "loss": 0.5707,
      "reward": 2.6460150480270386,
      "reward_std": 0.5471399649977684,
      "rewards/accuracy_reward": 0.755208358168602,
      "rewards/reasoning_steps_reward": 0.9774305671453476,
      "rewards/repetition_penalty_reward": -0.04625924210995436,
      "rewards/tag_count_reward": 0.9596354365348816,
      "step": 623
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 220.9322967529297,
      "epoch": 0.9358830146231721,
      "grad_norm": 3.902612646632179,
      "kl": 0.83935546875,
      "learning_rate": 2.4163390981474354e-07,
      "loss": 0.2083,
      "reward": 2.6754491925239563,
      "reward_std": 0.3636674992740154,
      "rewards/accuracy_reward": 0.7500000298023224,
      "rewards/reasoning_steps_reward": 0.9895833432674408,
      "rewards/repetition_penalty_reward": -0.041998728178441525,
      "rewards/tag_count_reward": 0.977864608168602,
      "step": 624
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 235.92187881469727,
      "epoch": 0.9373828271466067,
      "grad_norm": 31.877078305835827,
      "kl": 4.0859375,
      "learning_rate": 2.3030836230863108e-07,
      "loss": 0.728,
      "reward": 2.5619365572929382,
      "reward_std": 0.5274968221783638,
      "rewards/accuracy_reward": 0.677083358168602,
      "rewards/reasoning_steps_reward": 0.9809028059244156,
      "rewards/repetition_penalty_reward": -0.05568512622267008,
      "rewards/tag_count_reward": 0.9596354514360428,
      "step": 625
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 234.72917556762695,
      "epoch": 0.9388826396700413,
      "grad_norm": 5.128413845770082,
      "kl": 3.59765625,
      "learning_rate": 2.192515508722559e-07,
      "loss": 0.5811,
      "reward": 2.5278205275535583,
      "reward_std": 0.5462368726730347,
      "rewards/accuracy_reward": 0.6510416865348816,
      "rewards/reasoning_steps_reward": 0.9687500298023224,
      "rewards/repetition_penalty_reward": -0.04900246858596802,
      "rewards/tag_count_reward": 0.9570312649011612,
      "step": 626
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 217.0572967529297,
      "epoch": 0.9403824521934758,
      "grad_norm": 11.417578022441003,
      "kl": 3.625,
      "learning_rate": 2.08463779646646e-07,
      "loss": 0.6657,
      "reward": 2.617713212966919,
      "reward_std": 0.613451674580574,
      "rewards/accuracy_reward": 0.7500000149011612,
      "rewards/reasoning_steps_reward": 0.9652778208255768,
      "rewards/repetition_penalty_reward": -0.049387591890990734,
      "rewards/tag_count_reward": 0.9518229365348816,
      "step": 627
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 220.74480056762695,
      "epoch": 0.9418822647169104,
      "grad_norm": 13.927624029030772,
      "kl": 6.705078125,
      "learning_rate": 1.979453453723057e-07,
      "loss": 0.5758,
      "reward": 2.5704716444015503,
      "reward_std": 0.4331911653280258,
      "rewards/accuracy_reward": 0.6770833432674408,
      "rewards/reasoning_steps_reward": 0.9791666716337204,
      "rewards/repetition_penalty_reward": -0.03890332626178861,
      "rewards/tag_count_reward": 0.9531250298023224,
      "step": 628
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 230.04687881469727,
      "epoch": 0.943382077240345,
      "grad_norm": 2.359934750542652,
      "kl": 1.4609375,
      "learning_rate": 1.8769653738105797e-07,
      "loss": 0.1944,
      "reward": 2.650016129016876,
      "reward_std": 0.4671716019511223,
      "rewards/accuracy_reward": 0.7812500298023224,
      "rewards/reasoning_steps_reward": 0.9479166865348816,
      "rewards/repetition_penalty_reward": -0.05180681962519884,
      "rewards/tag_count_reward": 0.9726562649011612,
      "step": 629
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 241.45312881469727,
      "epoch": 0.9448818897637795,
      "grad_norm": 6.681812402325648,
      "kl": 4.34375,
      "learning_rate": 1.7771763758808403e-07,
      "loss": 0.7057,
      "reward": 2.5360206961631775,
      "reward_std": 0.6756476908922195,
      "rewards/accuracy_reward": 0.6875000298023224,
      "rewards/reasoning_steps_reward": 0.9704861342906952,
      "rewards/repetition_penalty_reward": -0.06727802660316229,
      "rewards/tag_count_reward": 0.9453125149011612,
      "step": 630
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 193.1041717529297,
      "epoch": 0.9463817022872141,
      "grad_norm": 4.5498762858817585,
      "kl": 2.009765625,
      "learning_rate": 1.6800892048416618e-07,
      "loss": 0.4715,
      "reward": 2.745138168334961,
      "reward_std": 0.4257928729057312,
      "rewards/accuracy_reward": 0.8177083730697632,
      "rewards/reasoning_steps_reward": 0.9878472238779068,
      "rewards/repetition_penalty_reward": -0.03828221885487437,
      "rewards/tag_count_reward": 0.977864608168602,
      "step": 631
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 224.62500762939453,
      "epoch": 0.9478815148106486,
      "grad_norm": 2.3610080061202314,
      "kl": 3.5078125,
      "learning_rate": 1.5857065312814058e-07,
      "loss": 0.786,
      "reward": 2.422154188156128,
      "reward_std": 0.622757077217102,
      "rewards/accuracy_reward": 0.5781250074505806,
      "rewards/reasoning_steps_reward": 0.956597238779068,
      "rewards/repetition_penalty_reward": -0.05657860916107893,
      "rewards/tag_count_reward": 0.9440104365348816,
      "step": 632
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 206.20312881469727,
      "epoch": 0.9493813273340832,
      "grad_norm": 14.033209063185351,
      "kl": 3.33984375,
      "learning_rate": 1.4940309513955088e-07,
      "loss": 0.782,
      "reward": 2.64290452003479,
      "reward_std": 0.5470430329442024,
      "rewards/accuracy_reward": 0.7447917014360428,
      "rewards/reasoning_steps_reward": 0.987847238779068,
      "rewards/repetition_penalty_reward": -0.048067858442664146,
      "rewards/tag_count_reward": 0.958333358168602,
      "step": 633
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 217.67709350585938,
      "epoch": 0.9508811398575178,
      "grad_norm": 1.7544974641972677,
      "kl": 2.810546875,
      "learning_rate": 1.405064986915028e-07,
      "loss": 0.5779,
      "reward": 2.556717336177826,
      "reward_std": 0.5784207582473755,
      "rewards/accuracy_reward": 0.6927083432674408,
      "rewards/reasoning_steps_reward": 0.9583333730697632,
      "rewards/repetition_penalty_reward": -0.051355627831071615,
      "rewards/tag_count_reward": 0.9570312649011612,
      "step": 634
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 232.2031364440918,
      "epoch": 0.9523809523809523,
      "grad_norm": 2.733833244509274,
      "kl": 2.9609375,
      "learning_rate": 1.3188110850373527e-07,
      "loss": 0.5971,
      "reward": 2.5860196948051453,
      "reward_std": 0.5797842293977737,
      "rewards/accuracy_reward": 0.7135416865348816,
      "rewards/reasoning_steps_reward": 0.970486119389534,
      "rewards/repetition_penalty_reward": -0.05503946170210838,
      "rewards/tag_count_reward": 0.9570312798023224,
      "step": 635
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 256.3697967529297,
      "epoch": 0.953880764904387,
      "grad_norm": 16.69375796765008,
      "kl": 4.93359375,
      "learning_rate": 1.2352716183588022e-07,
      "loss": 0.9024,
      "reward": 2.4972673654556274,
      "reward_std": 0.5852114260196686,
      "rewards/accuracy_reward": 0.6822916865348816,
      "rewards/reasoning_steps_reward": 0.9565972536802292,
      "rewards/repetition_penalty_reward": -0.0674027856439352,
      "rewards/tag_count_reward": 0.9257812649011612,
      "step": 636
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 237.43750381469727,
      "epoch": 0.9553805774278216,
      "grad_norm": 2.7912977723178796,
      "kl": 2.740234375,
      "learning_rate": 1.1544488848094338e-07,
      "loss": 0.4881,
      "reward": 2.582638919353485,
      "reward_std": 0.6020847111940384,
      "rewards/accuracy_reward": 0.7343750149011612,
      "rewards/reasoning_steps_reward": 0.9652778208255768,
      "rewards/repetition_penalty_reward": -0.05190982669591904,
      "rewards/tag_count_reward": 0.934895858168602,
      "step": 637
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 240.64063262939453,
      "epoch": 0.9568803899512561,
      "grad_norm": 6.968113723896311,
      "kl": 2.9375,
      "learning_rate": 1.0763451075897713e-07,
      "loss": 0.6358,
      "reward": 2.5640820264816284,
      "reward_std": 0.6016581207513809,
      "rewards/accuracy_reward": 0.6875000223517418,
      "rewards/reasoning_steps_reward": 0.9791666716337204,
      "rewards/repetition_penalty_reward": -0.0635221004486084,
      "rewards/tag_count_reward": 0.9609375298023224,
      "step": 638
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 226.46875762939453,
      "epoch": 0.9583802024746907,
      "grad_norm": 3.460380168793295,
      "kl": 3.59765625,
      "learning_rate": 1.0009624351097313e-07,
      "loss": 0.5328,
      "reward": 2.6055824160575867,
      "reward_std": 0.5846581794321537,
      "rewards/accuracy_reward": 0.7447916865348816,
      "rewards/reasoning_steps_reward": 0.9600694477558136,
      "rewards/repetition_penalty_reward": -0.05240369774401188,
      "rewards/tag_count_reward": 0.9531250149011612,
      "step": 639
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 232.04688262939453,
      "epoch": 0.9598800149981253,
      "grad_norm": 5.342162377565693,
      "kl": 2.919921875,
      "learning_rate": 9.283029409294263e-08,
      "loss": 0.5766,
      "reward": 2.615228056907654,
      "reward_std": 0.5253854840993881,
      "rewards/accuracy_reward": 0.7291667014360428,
      "rewards/reasoning_steps_reward": 0.972222238779068,
      "rewards/repetition_penalty_reward": -0.058817002922296524,
      "rewards/tag_count_reward": 0.9726562649011612,
      "step": 640
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 231.62500762939453,
      "epoch": 0.9613798275215598,
      "grad_norm": 15.144779529429583,
      "kl": 2.9951171875,
      "learning_rate": 8.583686237022105e-08,
      "loss": 0.6178,
      "reward": 2.459158480167389,
      "reward_std": 0.49489714950323105,
      "rewards/accuracy_reward": 0.567708358168602,
      "rewards/reasoning_steps_reward": 0.9791667014360428,
      "rewards/repetition_penalty_reward": -0.055164570454508066,
      "rewards/tag_count_reward": 0.9674479365348816,
      "step": 641
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 248.58334732055664,
      "epoch": 0.9628796400449944,
      "grad_norm": 7.8949182961312445,
      "kl": 3.7138671875,
      "learning_rate": 7.911614071196671e-08,
      "loss": 0.7138,
      "reward": 2.503873646259308,
      "reward_std": 0.4214767701923847,
      "rewards/accuracy_reward": 0.6562500149011612,
      "rewards/reasoning_steps_reward": 0.9670139104127884,
      "rewards/repetition_penalty_reward": -0.07251533772796392,
      "rewards/tag_count_reward": 0.953125,
      "step": 642
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 221.640625,
      "epoch": 0.9643794525684289,
      "grad_norm": 3.935827963006508,
      "kl": 2.220703125,
      "learning_rate": 7.266831398587082e-08,
      "loss": 0.3811,
      "reward": 2.5909196734428406,
      "reward_std": 0.5641670525074005,
      "rewards/accuracy_reward": 0.7031250149011612,
      "rewards/reasoning_steps_reward": 0.9791666865348816,
      "rewards/repetition_penalty_reward": -0.043195023201406,
      "rewards/tag_count_reward": 0.9518229514360428,
      "step": 643
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 249.91667938232422,
      "epoch": 0.9658792650918635,
      "grad_norm": 3.3366997354557224,
      "kl": 2.41796875,
      "learning_rate": 6.649355955306802e-08,
      "loss": 0.558,
      "reward": 2.6643757820129395,
      "reward_std": 0.4191970229148865,
      "rewards/accuracy_reward": 0.770833358168602,
      "rewards/reasoning_steps_reward": 0.9809028059244156,
      "rewards/repetition_penalty_reward": -0.05611046589910984,
      "rewards/tag_count_reward": 0.9687500298023224,
      "step": 644
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 232.72396850585938,
      "epoch": 0.9673790776152981,
      "grad_norm": 6.998025777128476,
      "kl": 2.767578125,
      "learning_rate": 6.059204726326373e-08,
      "loss": 0.6031,
      "reward": 2.2929863333702087,
      "reward_std": 0.6047337800264359,
      "rewards/accuracy_reward": 0.416666679084301,
      "rewards/reasoning_steps_reward": 0.973958358168602,
      "rewards/repetition_penalty_reward": -0.04815956763923168,
      "rewards/tag_count_reward": 0.9505208432674408,
      "step": 645
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 209.05208587646484,
      "epoch": 0.9688788901387326,
      "grad_norm": 7.463341518776408,
      "kl": 2.57421875,
      "learning_rate": 5.4963939450057846e-08,
      "loss": 0.5299,
      "reward": 2.7416443824768066,
      "reward_std": 0.513618029654026,
      "rewards/accuracy_reward": 0.8385416865348816,
      "rewards/reasoning_steps_reward": 0.9774305820465088,
      "rewards/repetition_penalty_reward": -0.03917174320667982,
      "rewards/tag_count_reward": 0.9648437798023224,
      "step": 646
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 230.57813262939453,
      "epoch": 0.9703787026621672,
      "grad_norm": 6.850765735428718,
      "kl": 1.62255859375,
      "learning_rate": 4.960939092648165e-08,
      "loss": 0.3969,
      "reward": 2.651937246322632,
      "reward_std": 0.4466264098882675,
      "rewards/accuracy_reward": 0.7343750149011612,
      "rewards/reasoning_steps_reward": 0.9947916865348816,
      "rewards/repetition_penalty_reward": -0.04728165362030268,
      "rewards/tag_count_reward": 0.970052108168602,
      "step": 647
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 244.28646087646484,
      "epoch": 0.9718785151856018,
      "grad_norm": 18.96391662542068,
      "kl": 3.640625,
      "learning_rate": 4.452854898073788e-08,
      "loss": 0.5956,
      "reward": 2.4319705963134766,
      "reward_std": 0.581806406378746,
      "rewards/accuracy_reward": 0.5885416716337204,
      "rewards/reasoning_steps_reward": 0.9565972983837128,
      "rewards/repetition_penalty_reward": -0.06368919461965561,
      "rewards/tag_count_reward": 0.950520858168602,
      "step": 648
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 228.5677146911621,
      "epoch": 0.9733783277090363,
      "grad_norm": 6.53111177098122,
      "kl": 2.421875,
      "learning_rate": 3.9721553372150665e-08,
      "loss": 0.4821,
      "reward": 2.584895968437195,
      "reward_std": 0.510861761868,
      "rewards/accuracy_reward": 0.697916679084301,
      "rewards/reasoning_steps_reward": 0.9670139104127884,
      "rewards/repetition_penalty_reward": -0.048784732818603516,
      "rewards/tag_count_reward": 0.9687500149011612,
      "step": 649
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 248.70834350585938,
      "epoch": 0.974878140232471,
      "grad_norm": 8.617456792461612,
      "kl": 3.447265625,
      "learning_rate": 3.5188536327318554e-08,
      "loss": 0.7325,
      "reward": 2.51455420255661,
      "reward_std": 0.6581666991114616,
      "rewards/accuracy_reward": 0.6614583432674408,
      "rewards/reasoning_steps_reward": 0.9652777761220932,
      "rewards/repetition_penalty_reward": -0.05879658181220293,
      "rewards/tag_count_reward": 0.946614608168602,
      "step": 650
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 219.52083587646484,
      "epoch": 0.9763779527559056,
      "grad_norm": 6.114393105789686,
      "kl": 2.314453125,
      "learning_rate": 3.092962253648302e-08,
      "loss": 0.4754,
      "reward": 2.638841927051544,
      "reward_std": 0.5403295606374741,
      "rewards/accuracy_reward": 0.7500000149011612,
      "rewards/reasoning_steps_reward": 0.9739583432674408,
      "rewards/repetition_penalty_reward": -0.044752005487680435,
      "rewards/tag_count_reward": 0.9596354365348816,
      "step": 651
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 222.9791717529297,
      "epoch": 0.9778777652793401,
      "grad_norm": 4.664735172295086,
      "kl": 4.345703125,
      "learning_rate": 2.694492915009006e-08,
      "loss": 0.5058,
      "reward": 2.429598391056061,
      "reward_std": 0.4925737604498863,
      "rewards/accuracy_reward": 0.5468750149011612,
      "rewards/reasoning_steps_reward": 0.9791666865348816,
      "rewards/repetition_penalty_reward": -0.0482663013972342,
      "rewards/tag_count_reward": 0.9518229365348816,
      "step": 652
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 262.4739685058594,
      "epoch": 0.9793775778027747,
      "grad_norm": 9.946082586920667,
      "kl": 3.6181640625,
      "learning_rate": 2.3234565775575034e-08,
      "loss": 0.7296,
      "reward": 2.446805000305176,
      "reward_std": 0.5700404495000839,
      "rewards/accuracy_reward": 0.5989583432674408,
      "rewards/reasoning_steps_reward": 0.9756944626569748,
      "rewards/repetition_penalty_reward": -0.07316032983362675,
      "rewards/tag_count_reward": 0.9453125298023224,
      "step": 653
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 213.90625381469727,
      "epoch": 0.9808773903262092,
      "grad_norm": 7.538573623853772,
      "kl": 2.619140625,
      "learning_rate": 1.9798634474345048e-08,
      "loss": 0.5537,
      "reward": 2.517378091812134,
      "reward_std": 0.41347331553697586,
      "rewards/accuracy_reward": 0.5989583507180214,
      "rewards/reasoning_steps_reward": 0.9895833432674408,
      "rewards/repetition_penalty_reward": -0.043819915503263474,
      "rewards/tag_count_reward": 0.9726562649011612,
      "step": 654
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 222.4791717529297,
      "epoch": 0.9823772028496438,
      "grad_norm": 1.4494224041136212,
      "kl": 1.38916015625,
      "learning_rate": 1.6637229758970087e-08,
      "loss": 0.2744,
      "reward": 2.6669066548347473,
      "reward_std": 0.3448048084974289,
      "rewards/accuracy_reward": 0.7447916865348816,
      "rewards/reasoning_steps_reward": 0.9878472238779068,
      "rewards/repetition_penalty_reward": -0.04229486454278231,
      "rewards/tag_count_reward": 0.9765625149011612,
      "step": 655
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 241.97396850585938,
      "epoch": 0.9838770153730784,
      "grad_norm": 5.752377864686793,
      "kl": 1.896484375,
      "learning_rate": 1.3750438590586223e-08,
      "loss": 0.4106,
      "reward": 2.751551926136017,
      "reward_std": 0.3733288496732712,
      "rewards/accuracy_reward": 0.8437500149011612,
      "rewards/reasoning_steps_reward": 0.9878472536802292,
      "rewards/repetition_penalty_reward": -0.055305857211351395,
      "rewards/tag_count_reward": 0.9752604365348816,
      "step": 656
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 233.6093864440918,
      "epoch": 0.9853768278965129,
      "grad_norm": 1.7146871873625331,
      "kl": 2.14794921875,
      "learning_rate": 1.1138340376501966e-08,
      "loss": 0.2227,
      "reward": 2.6160547137260437,
      "reward_std": 0.5579579994082451,
      "rewards/accuracy_reward": 0.7343750149011612,
      "rewards/reasoning_steps_reward": 0.9670139104127884,
      "rewards/repetition_penalty_reward": -0.05538627551868558,
      "rewards/tag_count_reward": 0.9700520932674408,
      "step": 657
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 221.94271087646484,
      "epoch": 0.9868766404199475,
      "grad_norm": 9.514836835014375,
      "kl": 3.48046875,
      "learning_rate": 8.801006968012227e-09,
      "loss": 0.6615,
      "reward": 2.577666759490967,
      "reward_std": 0.4780506566166878,
      "rewards/accuracy_reward": 0.6875000149011612,
      "rewards/reasoning_steps_reward": 0.9809028059244156,
      "rewards/repetition_penalty_reward": -0.05167360603809357,
      "rewards/tag_count_reward": 0.9609375149011612,
      "step": 658
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 194.54687881469727,
      "epoch": 0.9883764529433821,
      "grad_norm": 7.048358144573533,
      "kl": 2.072265625,
      "learning_rate": 6.738502658426571e-09,
      "loss": 0.4597,
      "reward": 2.7473713755607605,
      "reward_std": 0.4053123965859413,
      "rewards/accuracy_reward": 0.8281250149011612,
      "rewards/reasoning_steps_reward": 0.9861111342906952,
      "rewards/repetition_penalty_reward": -0.035614716820418835,
      "rewards/tag_count_reward": 0.9687500298023224,
      "step": 659
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 197.7760467529297,
      "epoch": 0.9898762654668166,
      "grad_norm": 7.604534130768124,
      "kl": 1.7197265625,
      "learning_rate": 4.950884181295079e-09,
      "loss": 0.2393,
      "reward": 2.58710515499115,
      "reward_std": 0.4081360250711441,
      "rewards/accuracy_reward": 0.661458358168602,
      "rewards/reasoning_steps_reward": 0.986111119389534,
      "rewards/repetition_penalty_reward": -0.04353731218725443,
      "rewards/tag_count_reward": 0.9830729514360428,
      "step": 660
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 249.40105056762695,
      "epoch": 0.9913760779902512,
      "grad_norm": 7.864265781403792,
      "kl": 4.31640625,
      "learning_rate": 3.4382007088518134e-09,
      "loss": 0.7781,
      "reward": 2.3653112053871155,
      "reward_std": 0.5962060615420341,
      "rewards/accuracy_reward": 0.5468750149011612,
      "rewards/reasoning_steps_reward": 0.9479167014360428,
      "rewards/repetition_penalty_reward": -0.06047023739665747,
      "rewards/tag_count_reward": 0.930989608168602,
      "step": 661
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 230.3489646911621,
      "epoch": 0.9928758905136857,
      "grad_norm": 3.5888298694634995,
      "kl": 3.0,
      "learning_rate": 2.200493850662566e-09,
      "loss": 0.5777,
      "reward": 2.563965141773224,
      "reward_std": 0.5070576220750809,
      "rewards/accuracy_reward": 0.6822916716337204,
      "rewards/reasoning_steps_reward": 0.9826389104127884,
      "rewards/repetition_penalty_reward": -0.057996807619929314,
      "rewards/tag_count_reward": 0.9570312649011612,
      "step": 662
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 223.26041793823242,
      "epoch": 0.9943757030371203,
      "grad_norm": 4.0715790482727705,
      "kl": 2.197265625,
      "learning_rate": 1.2377976524746705e-09,
      "loss": 0.2855,
      "reward": 2.672169864177704,
      "reward_std": 0.48223114758729935,
      "rewards/accuracy_reward": 0.7916666865348816,
      "rewards/reasoning_steps_reward": 0.9687500447034836,
      "rewards/repetition_penalty_reward": -0.04527819249778986,
      "rewards/tag_count_reward": 0.9570312649011612,
      "step": 663
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 246.0677146911621,
      "epoch": 0.995875515560555,
      "grad_norm": 3.596922089624617,
      "kl": 2.80078125,
      "learning_rate": 5.501385952888516e-10,
      "loss": 0.8216,
      "reward": 2.5335150957107544,
      "reward_std": 0.5313072800636292,
      "rewards/accuracy_reward": 0.6302083507180214,
      "rewards/reasoning_steps_reward": 0.9947916716337204,
      "rewards/repetition_penalty_reward": -0.04981834441423416,
      "rewards/tag_count_reward": 0.9583333432674408,
      "step": 664
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 218.2447967529297,
      "epoch": 0.9973753280839895,
      "grad_norm": 3.9248590354460755,
      "kl": 1.875,
      "learning_rate": 1.375355946242607e-10,
      "loss": 0.3974,
      "reward": 2.5833805203437805,
      "reward_std": 0.4426957219839096,
      "rewards/accuracy_reward": 0.6718750298023224,
      "rewards/reasoning_steps_reward": 0.9809027910232544,
      "rewards/repetition_penalty_reward": -0.03944949712604284,
      "rewards/tag_count_reward": 0.970052108168602,
      "step": 665
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 223.3671875,
      "epoch": 0.9988751406074241,
      "grad_norm": 7.299493407835698,
      "kl": 2.71875,
      "learning_rate": 0.0,
      "loss": 0.5344,
      "reward": 2.6338536143302917,
      "reward_std": 0.6461013555526733,
      "rewards/accuracy_reward": 0.7656250149011612,
      "rewards/reasoning_steps_reward": 0.963541716337204,
      "rewards/repetition_penalty_reward": -0.045833950862288475,
      "rewards/tag_count_reward": 0.950520858168602,
      "step": 666
    },
    {
      "epoch": 0.9988751406074241,
      "step": 666,
      "total_flos": 0.0,
      "train_loss": 14.041643001426444,
      "train_runtime": 23054.463,
      "train_samples_per_second": 0.347,
      "train_steps_per_second": 0.029
    }
  ],
  "logging_steps": 1,
  "max_steps": 666,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 200,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}