{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.1713796058269066,
  "eval_steps": 500,
  "global_step": 100,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "completion_length": 2842.9896545410156,
      "epoch": 0.001713796058269066,
      "grad_norm": 0.16336098313331604,
      "kl": 0.0,
      "learning_rate": 1e-07,
      "loss": 0.0,
      "reward": 0.0586470989510417,
      "reward_std": 0.1930437758564949,
      "rewards/cosine_scaled_reward": -0.0629267618060112,
      "rewards/format_reward": 0.3541666716337204,
      "step": 1
    },
    {
      "completion_length": 3090.7500610351562,
      "epoch": 0.003427592116538132,
      "grad_norm": 0.15128685534000397,
      "kl": 0.0,
      "learning_rate": 2e-07,
      "loss": -0.0,
      "reward": 0.013502426445484161,
      "reward_std": 0.14039192721247673,
      "rewards/cosine_scaled_reward": -0.10344066098332405,
      "rewards/format_reward": 0.260416679084301,
      "step": 2
    },
    {
      "completion_length": 2600.1875610351562,
      "epoch": 0.005141388174807198,
      "grad_norm": 0.1910804808139801,
      "kl": 3.987550735473633e-05,
      "learning_rate": 3e-07,
      "loss": 0.0,
      "reward": 0.037791766226291656,
      "reward_std": 0.1596581730991602,
      "rewards/cosine_scaled_reward": -0.1509031739551574,
      "rewards/format_reward": 0.4479166865348816,
      "step": 3
    },
    {
      "completion_length": 3076.8438110351562,
      "epoch": 0.006855184233076264,
      "grad_norm": 0.15531130135059357,
      "kl": 4.374980926513672e-05,
      "learning_rate": 4e-07,
      "loss": 0.0,
      "reward": -0.010211404645815492,
      "reward_std": 0.18583546206355095,
      "rewards/cosine_scaled_reward": -0.19252582639455795,
      "rewards/format_reward": 0.3437500037252903,
      "step": 4
    },
    {
      "completion_length": 3045.604248046875,
      "epoch": 0.00856898029134533,
      "grad_norm": 0.11871814727783203,
      "kl": 2.682209014892578e-05,
      "learning_rate": 5e-07,
      "loss": 0.0,
      "reward": 0.0951396506279707,
      "reward_std": 0.17187905684113503,
      "rewards/cosine_scaled_reward": -0.026736530475318432,
      "rewards/format_reward": 0.4270833358168602,
      "step": 5
    },
    {
      "completion_length": 2959.3438720703125,
      "epoch": 0.010282776349614395,
      "grad_norm": 0.13790491223335266,
      "kl": 3.451108932495117e-05,
      "learning_rate": 6e-07,
      "loss": 0.0,
      "reward": 0.11859842715784907,
      "reward_std": 0.17868740856647491,
      "rewards/cosine_scaled_reward": 0.010096830315887928,
      "rewards/format_reward": 0.447916679084301,
      "step": 6
    },
    {
      "completion_length": 2884.6771850585938,
      "epoch": 0.011996572407883462,
      "grad_norm": 0.1611727625131607,
      "kl": 2.6971101760864258e-05,
      "learning_rate": 7e-07,
      "loss": 0.0,
      "reward": 0.03227711468935013,
      "reward_std": 0.19281568750739098,
      "rewards/cosine_scaled_reward": -0.12487578392028809,
      "rewards/format_reward": 0.3750000149011612,
      "step": 7
    },
    {
      "completion_length": 2835.6666870117188,
      "epoch": 0.013710368466152529,
      "grad_norm": 0.1475144475698471,
      "kl": 3.361701965332031e-05,
      "learning_rate": 8e-07,
      "loss": 0.0,
      "reward": 0.045130204409360886,
      "reward_std": 0.18995673581957817,
      "rewards/cosine_scaled_reward": -0.13529739249497652,
      "rewards/format_reward": 0.44791667722165585,
      "step": 8
    },
    {
      "completion_length": 3167.3541870117188,
      "epoch": 0.015424164524421594,
      "grad_norm": 0.13677193224430084,
      "kl": 2.440810203552246e-05,
      "learning_rate": 9e-07,
      "loss": 0.0,
      "reward": 0.020484148059040308,
      "reward_std": 0.16632195003330708,
      "rewards/cosine_scaled_reward": -0.11485888389870524,
      "rewards/format_reward": 0.3125000111758709,
      "step": 9
    },
    {
      "completion_length": 2629.4584350585938,
      "epoch": 0.01713796058269066,
      "grad_norm": 0.14359042048454285,
      "kl": 2.0831823348999023e-05,
      "learning_rate": 1e-06,
      "loss": 0.0,
      "reward": 0.1312947366386652,
      "reward_std": 0.1554134227335453,
      "rewards/cosine_scaled_reward": 0.009967727586627007,
      "rewards/format_reward": 0.5000000074505806,
      "step": 10
    },
    {
      "completion_length": 2906.1875610351562,
      "epoch": 0.018851756640959727,
      "grad_norm": 0.20071381330490112,
      "kl": 3.3527612686157227e-05,
      "learning_rate": 9.997258721585931e-07,
      "loss": 0.0,
      "reward": 0.033789366483688354,
      "reward_std": 0.19262398406863213,
      "rewards/cosine_scaled_reward": -0.08468496927525848,
      "rewards/format_reward": 0.3020833460614085,
      "step": 11
    },
    {
      "completion_length": 2942.4168090820312,
      "epoch": 0.02056555269922879,
      "grad_norm": 0.15123063325881958,
      "kl": 2.7477741241455078e-05,
      "learning_rate": 9.989038226169207e-07,
      "loss": 0.0,
      "reward": 0.052781688049435616,
      "reward_std": 0.19968328438699245,
      "rewards/cosine_scaled_reward": -0.08891183231025934,
      "rewards/format_reward": 0.3854166865348816,
      "step": 12
    },
    {
      "completion_length": 2924.7501220703125,
      "epoch": 0.022279348757497857,
      "grad_norm": 0.1488049179315567,
      "kl": 1.4431774616241455e-05,
      "learning_rate": 9.975348529157229e-07,
      "loss": 0.0,
      "reward": 0.1416405034251511,
      "reward_std": 0.20871220529079437,
      "rewards/cosine_scaled_reward": 0.05599783733487129,
      "rewards/format_reward": 0.4479166865348816,
      "step": 13
    },
    {
      "completion_length": 2709.6250915527344,
      "epoch": 0.023993144815766924,
      "grad_norm": 0.18846973776817322,
      "kl": 3.9637088775634766e-05,
      "learning_rate": 9.956206309337066e-07,
      "loss": 0.0,
      "reward": 0.06054589990526438,
      "reward_std": 0.1687684953212738,
      "rewards/cosine_scaled_reward": -0.10551007091999054,
      "rewards/format_reward": 0.4479166865348816,
      "step": 14
    },
    {
      "completion_length": 1950.8646545410156,
      "epoch": 0.02570694087403599,
      "grad_norm": 0.18432216346263885,
      "kl": 4.6759843826293945e-05,
      "learning_rate": 9.931634888554935e-07,
      "loss": 0.0,
      "reward": 0.23200466111302376,
      "reward_std": 0.2355622537434101,
      "rewards/cosine_scaled_reward": 0.06849630922079086,
      "rewards/format_reward": 0.7708333432674408,
      "step": 15
    },
    {
      "completion_length": 3007.6875,
      "epoch": 0.027420736932305057,
      "grad_norm": 0.15038132667541504,
      "kl": 3.84598970413208e-05,
      "learning_rate": 9.901664203302124e-07,
      "loss": 0.0,
      "reward": 0.06949906051158905,
      "reward_std": 0.21871956065297127,
      "rewards/cosine_scaled_reward": -0.07155923545360565,
      "rewards/format_reward": 0.416666679084301,
      "step": 16
    },
    {
      "completion_length": 2888.8438110351562,
      "epoch": 0.02913453299057412,
      "grad_norm": 0.17018933594226837,
      "kl": 5.860626697540283e-05,
      "learning_rate": 9.866330768241983e-07,
      "loss": 0.0,
      "reward": 0.08268930693157017,
      "reward_std": 0.2000725194811821,
      "rewards/cosine_scaled_reward": -0.03420639084652066,
      "rewards/format_reward": 0.3958333469927311,
      "step": 17
    },
    {
      "completion_length": 2994.1666870117188,
      "epoch": 0.030848329048843187,
      "grad_norm": 0.14627192914485931,
      "kl": 3.9910897612571716e-05,
      "learning_rate": 9.825677631722435e-07,
      "loss": 0.0,
      "reward": 0.1066256552003324,
      "reward_std": 0.1765430960804224,
      "rewards/cosine_scaled_reward": -0.012849166989326477,
      "rewards/format_reward": 0.447916679084301,
      "step": 18
    },
    {
      "completion_length": 2986.885498046875,
      "epoch": 0.032562125107112254,
      "grad_norm": 0.11996503919363022,
      "kl": 7.943809032440186e-05,
      "learning_rate": 9.779754323328192e-07,
      "loss": 0.0,
      "reward": 0.037288530729711056,
      "reward_std": 0.1901964209973812,
      "rewards/cosine_scaled_reward": -0.08846686349716038,
      "rewards/format_reward": 0.3229166753590107,
      "step": 19
    },
    {
      "completion_length": 3068.0000610351562,
      "epoch": 0.03427592116538132,
      "grad_norm": 0.15175943076610565,
      "kl": 0.00019240379333496094,
      "learning_rate": 9.728616793536587e-07,
      "loss": 0.0,
      "reward": 0.03987372852861881,
      "reward_std": 0.18592991679906845,
      "rewards/cosine_scaled_reward": -0.11358156427741051,
      "rewards/format_reward": 0.3854166828095913,
      "step": 20
    },
    {
      "completion_length": 3086.6771240234375,
      "epoch": 0.03598971722365039,
      "grad_norm": 0.1517907977104187,
      "kl": 0.00036978721618652344,
      "learning_rate": 9.672327345550543e-07,
      "loss": 0.0,
      "reward": 0.03412920422852039,
      "reward_std": 0.1651427522301674,
      "rewards/cosine_scaled_reward": -0.09963592886924744,
      "rewards/format_reward": 0.33333334419876337,
      "step": 21
    },
    {
      "completion_length": 3277.1458740234375,
      "epoch": 0.037703513281919454,
      "grad_norm": 0.11310048401355743,
      "kl": 0.00019124150276184082,
      "learning_rate": 9.610954559391704e-07,
      "loss": 0.0,
      "reward": -0.0021694526076316833,
      "reward_std": 0.1807849071919918,
      "rewards/cosine_scaled_reward": -0.1448658797889948,
      "rewards/format_reward": 0.2812500074505806,
      "step": 22
    },
    {
      "completion_length": 2485.1875915527344,
      "epoch": 0.03941730934018852,
      "grad_norm": 0.21775458753108978,
      "kl": 0.0002810359001159668,
      "learning_rate": 9.54457320834625e-07,
      "loss": 0.0,
      "reward": 0.1259636478498578,
      "reward_std": 0.21954844892024994,
      "rewards/cosine_scaled_reward": 0.0026862788945436478,
      "rewards/format_reward": 0.4895833395421505,
      "step": 23
    },
    {
      "completion_length": 3387.8646240234375,
      "epoch": 0.04113110539845758,
      "grad_norm": 0.11246265470981598,
      "kl": 0.00029206275939941406,
      "learning_rate": 9.473264167865171e-07,
      "loss": 0.0,
      "reward": -0.0444450975919608,
      "reward_std": 0.19019928015768528,
      "rewards/cosine_scaled_reward": -0.17542381957173347,
      "rewards/format_reward": 0.1770833358168602,
      "step": 24
    },
    {
      "completion_length": 3383.3229370117188,
      "epoch": 0.04284490145672665,
      "grad_norm": 0.11982468515634537,
      "kl": 0.00018139183521270752,
      "learning_rate": 9.397114317029974e-07,
      "loss": 0.0,
      "reward": -0.08237307704985142,
      "reward_std": 0.10484589636325836,
      "rewards/cosine_scaled_reward": -0.2248186320066452,
      "rewards/format_reward": 0.1250000074505806,
      "step": 25
    },
    {
      "completion_length": 2839.0625610351562,
      "epoch": 0.044558697514995714,
      "grad_norm": 0.1347901076078415,
      "kl": 0.0003955364227294922,
      "learning_rate": 9.316216432703916e-07,
      "loss": 0.0,
      "reward": 0.11097948951646686,
      "reward_std": 0.18024785444140434,
      "rewards/cosine_scaled_reward": 0.0011768988333642483,
      "rewards/format_reward": 0.4375000111758709,
      "step": 26
    },
    {
      "completion_length": 2636.5729370117188,
      "epoch": 0.04627249357326478,
      "grad_norm": 0.13496308028697968,
      "kl": 0.0006954669952392578,
      "learning_rate": 9.230669076497687e-07,
      "loss": 0.0,
      "reward": 0.05549928080290556,
      "reward_std": 0.14237392507493496,
      "rewards/cosine_scaled_reward": -0.12638001888990402,
      "rewards/format_reward": 0.4687500111758709,
      "step": 27
    },
    {
      "completion_length": 2851.2396240234375,
      "epoch": 0.04798628963153385,
      "grad_norm": 0.16802968084812164,
      "kl": 0.00039768218994140625,
      "learning_rate": 9.140576474687263e-07,
      "loss": 0.0,
      "reward": 0.005757941631600261,
      "reward_std": 0.16327618807554245,
      "rewards/cosine_scaled_reward": -0.18160032108426094,
      "rewards/format_reward": 0.385416679084301,
      "step": 28
    },
    {
      "completion_length": 2759.5208740234375,
      "epoch": 0.049700085689802914,
      "grad_norm": 0.1442675143480301,
      "kl": 0.0005460977554321289,
      "learning_rate": 9.046048391230247e-07,
      "loss": 0.0,
      "reward": 0.09482090920209885,
      "reward_std": 0.16877998411655426,
      "rewards/cosine_scaled_reward": -0.010849054902791977,
      "rewards/format_reward": 0.3958333432674408,
      "step": 29
    },
    {
      "completion_length": 3377.59375,
      "epoch": 0.05141388174807198,
      "grad_norm": 0.12018371373414993,
      "kl": 0.0006362795829772949,
      "learning_rate": 8.9471999940354e-07,
      "loss": 0.0,
      "reward": 0.028713669627904892,
      "reward_std": 0.18587654829025269,
      "rewards/cosine_scaled_reward": -0.07821773737668991,
      "rewards/format_reward": 0.2708333358168602,
      "step": 30
    },
    {
      "completion_length": 3138.104248046875,
      "epoch": 0.05312767780634105,
      "grad_norm": 0.1430775374174118,
      "kl": 0.0004627704620361328,
      "learning_rate": 8.844151714648274e-07,
      "loss": 0.0,
      "reward": -0.008235585875809193,
      "reward_std": 0.14394641481339931,
      "rewards/cosine_scaled_reward": -0.13156688585877419,
      "rewards/format_reward": 0.22916666977107525,
      "step": 31
    },
    {
      "completion_length": 2630.5938720703125,
      "epoch": 0.054841473864610114,
      "grad_norm": 0.17797145247459412,
      "kl": 0.0012159347534179688,
      "learning_rate": 8.737029101523929e-07,
      "loss": 0.0,
      "reward": 0.13316596718505025,
      "reward_std": 0.22369278594851494,
      "rewards/cosine_scaled_reward": -0.002324562519788742,
      "rewards/format_reward": 0.5312500149011612,
      "step": 32
    },
    {
      "completion_length": 2592.5000610351562,
      "epoch": 0.056555269922879174,
      "grad_norm": 0.15806646645069122,
      "kl": 0.0012989044189453125,
      "learning_rate": 8.625962667065487e-07,
      "loss": 0.0001,
      "reward": 0.11066696606576443,
      "reward_std": 0.14327863790094852,
      "rewards/cosine_scaled_reward": -0.010827179066836834,
      "rewards/format_reward": 0.4583333432674408,
      "step": 33
    },
    {
      "completion_length": 2626.5313110351562,
      "epoch": 0.05826906598114824,
      "grad_norm": 0.1691393107175827,
      "kl": 0.0020818710327148438,
      "learning_rate": 8.511087728614862e-07,
      "loss": 0.0001,
      "reward": 0.09107339568436146,
      "reward_std": 0.18131284788250923,
      "rewards/cosine_scaled_reward": -0.04873488505836576,
      "rewards/format_reward": 0.4583333507180214,
      "step": 34
    },
    {
      "completion_length": 3051.9896850585938,
      "epoch": 0.05998286203941731,
      "grad_norm": 0.17819511890411377,
      "kl": 0.0019023418426513672,
      "learning_rate": 8.392544243589427e-07,
      "loss": 0.0001,
      "reward": 0.10884217114653438,
      "reward_std": 0.23975158482789993,
      "rewards/cosine_scaled_reward": 0.007055521011352539,
      "rewards/format_reward": 0.416666679084301,
      "step": 35
    },
    {
      "completion_length": 2809.8646240234375,
      "epoch": 0.061696658097686374,
      "grad_norm": 0.14143270254135132,
      "kl": 0.0004401206970214844,
      "learning_rate": 8.270476638965461e-07,
      "loss": 0.0,
      "reward": 0.15133970603346825,
      "reward_std": 0.25536274164915085,
      "rewards/cosine_scaled_reward": 0.04777143523097038,
      "rewards/format_reward": 0.5000000149011612,
      "step": 36
    },
    {
      "completion_length": 3153.0938110351562,
      "epoch": 0.06341045415595545,
      "grad_norm": 0.12571778893470764,
      "kl": 0.0004382133483886719,
      "learning_rate": 8.145033635316128e-07,
      "loss": 0.0,
      "reward": 0.020705484319478273,
      "reward_std": 0.17584117874503136,
      "rewards/cosine_scaled_reward": -0.13113276287913322,
      "rewards/format_reward": 0.3437500074505806,
      "step": 37
    },
    {
      "completion_length": 3118.322998046875,
      "epoch": 0.06512425021422451,
      "grad_norm": 0.11900849640369415,
      "kl": 0.00022101402282714844,
      "learning_rate": 8.01636806561836e-07,
      "loss": 0.0,
      "reward": 0.07037997618317604,
      "reward_std": 0.2210748754441738,
      "rewards/cosine_scaled_reward": -0.028155604377388954,
      "rewards/format_reward": 0.33333333395421505,
      "step": 38
    },
    {
      "completion_length": 2472.3646240234375,
      "epoch": 0.06683804627249357,
      "grad_norm": 0.1329745203256607,
      "kl": 0.0012660026550292969,
      "learning_rate": 7.884636689049422e-07,
      "loss": 0.0001,
      "reward": 0.11453905934467912,
      "reward_std": 0.20037821307778358,
      "rewards/cosine_scaled_reward": -0.056616831570863724,
      "rewards/format_reward": 0.5625000037252903,
      "step": 39
    },
    {
      "completion_length": 3018.072998046875,
      "epoch": 0.06855184233076264,
      "grad_norm": 0.11555609107017517,
      "kl": 0.00041091442108154297,
      "learning_rate": 7.75e-07,
      "loss": 0.0,
      "reward": 0.01033751294016838,
      "reward_std": 0.1491969907656312,
      "rewards/cosine_scaled_reward": -0.13083185872528702,
      "rewards/format_reward": 0.3020833386108279,
      "step": 40
    },
    {
      "completion_length": 2939.229248046875,
      "epoch": 0.0702656383890317,
      "grad_norm": 0.12907516956329346,
      "kl": 0.0007364749908447266,
      "learning_rate": 7.612622032536507e-07,
      "loss": 0.0,
      "reward": 0.12070174579275772,
      "reward_std": 0.2728600651025772,
      "rewards/cosine_scaled_reward": -0.001188491762150079,
      "rewards/format_reward": 0.479166679084301,
      "step": 41
    },
    {
      "completion_length": 2293.854248046875,
      "epoch": 0.07197943444730077,
      "grad_norm": 0.15430894494056702,
      "kl": 0.001895904541015625,
      "learning_rate": 7.472670160550848e-07,
      "loss": 0.0001,
      "reward": 0.1563777169212699,
      "reward_std": 0.2286355346441269,
      "rewards/cosine_scaled_reward": -0.015645149163901806,
      "rewards/format_reward": 0.6458333432674408,
      "step": 42
    },
    {
      "completion_length": 2741.0833740234375,
      "epoch": 0.07369323050556983,
      "grad_norm": 0.11995150148868561,
      "kl": 0.0004520416259765625,
      "learning_rate": 7.330314893841101e-07,
      "loss": 0.0,
      "reward": 0.13986177369952202,
      "reward_std": 0.22300692275166512,
      "rewards/cosine_scaled_reward": 0.05043928697705269,
      "rewards/format_reward": 0.4479166753590107,
      "step": 43
    },
    {
      "completion_length": 2305.854217529297,
      "epoch": 0.07540702656383891,
      "grad_norm": 0.21461769938468933,
      "kl": 0.0007772445678710938,
      "learning_rate": 7.185729670371604e-07,
      "loss": 0.0,
      "reward": 0.07035261578857899,
      "reward_std": 0.13939796388149261,
      "rewards/cosine_scaled_reward": -0.09040920436382294,
      "rewards/format_reward": 0.4583333386108279,
      "step": 44
    },
    {
      "completion_length": 2846.8958435058594,
      "epoch": 0.07712082262210797,
      "grad_norm": 0.13480690121650696,
      "kl": 0.0009918212890625,
      "learning_rate": 7.039090644965509e-07,
      "loss": 0.0,
      "reward": -0.018673544749617577,
      "reward_std": 0.16758902743458748,
      "rewards/cosine_scaled_reward": -0.20985213853418827,
      "rewards/format_reward": 0.3437500149011612,
      "step": 45
    },
    {
      "completion_length": 2584.1563110351562,
      "epoch": 0.07883461868037704,
      "grad_norm": 0.17533442378044128,
      "kl": 0.0022344589233398438,
      "learning_rate": 6.890576474687263e-07,
      "loss": 0.0001,
      "reward": 0.022231539245694876,
      "reward_std": 0.16662119142711163,
      "rewards/cosine_scaled_reward": -0.17729704454541206,
      "rewards/format_reward": 0.4375000223517418,
      "step": 46
    },
    {
      "completion_length": 3102.354248046875,
      "epoch": 0.0805484147386461,
      "grad_norm": 0.1247975304722786,
      "kl": 0.0013914108276367188,
      "learning_rate": 6.740368101176495e-07,
      "loss": 0.0001,
      "reward": -0.00813557580113411,
      "reward_std": 0.1496633905917406,
      "rewards/cosine_scaled_reward": -0.17180366069078445,
      "rewards/format_reward": 0.31250000838190317,
      "step": 47
    },
    {
      "completion_length": 2603.8438415527344,
      "epoch": 0.08226221079691516,
      "grad_norm": 0.16067557036876678,
      "kl": 0.0016384124755859375,
      "learning_rate": 6.588648530198504e-07,
      "loss": 0.0001,
      "reward": 0.11067175259813666,
      "reward_std": 0.179900873452425,
      "rewards/cosine_scaled_reward": -0.013014900498092175,
      "rewards/format_reward": 0.4583333507180214,
      "step": 48
    },
    {
      "completion_length": 3373.0313110351562,
      "epoch": 0.08397600685518423,
      "grad_norm": 0.10881118476390839,
      "kl": 0.0002281665802001953,
      "learning_rate": 6.435602608679916e-07,
      "loss": 0.0,
      "reward": 0.00016227364540100098,
      "reward_std": 0.14608008973300457,
      "rewards/cosine_scaled_reward": -0.10383528657257557,
      "rewards/format_reward": 0.2083333358168602,
      "step": 49
    },
    {
      "completion_length": 3033.0208740234375,
      "epoch": 0.0856898029134533,
      "grad_norm": 0.13846911489963531,
      "kl": 0.0012478828430175781,
      "learning_rate": 6.281416799501187e-07,
      "loss": 0.0001,
      "reward": 0.06065037357620895,
      "reward_std": 0.17429610714316368,
      "rewards/cosine_scaled_reward": -0.05210728943347931,
      "rewards/format_reward": 0.3437500074505806,
      "step": 50
    },
    {
      "completion_length": 2933.4793090820312,
      "epoch": 0.08740359897172237,
      "grad_norm": 0.13350166380405426,
      "kl": 0.00033092498779296875,
      "learning_rate": 6.126278954320294e-07,
      "loss": 0.0,
      "reward": -0.01886086305603385,
      "reward_std": 0.13372720032930374,
      "rewards/cosine_scaled_reward": -0.23485597968101501,
      "rewards/format_reward": 0.3958333507180214,
      "step": 51
    },
    {
      "completion_length": 3236.791748046875,
      "epoch": 0.08911739502999143,
      "grad_norm": 0.10703922808170319,
      "kl": 0.0003795623779296875,
      "learning_rate": 5.97037808470444e-07,
      "loss": 0.0,
      "reward": 0.0947269550524652,
      "reward_std": 0.2209848240017891,
      "rewards/cosine_scaled_reward": 0.004282929934561253,
      "rewards/format_reward": 0.3645833432674408,
      "step": 52
    },
    {
      "completion_length": 2535.1875610351562,
      "epoch": 0.0908311910882605,
      "grad_norm": 0.19642022252082825,
      "kl": 0.0011463165283203125,
      "learning_rate": 5.813904131848564e-07,
      "loss": 0.0,
      "reward": 0.074245510622859,
      "reward_std": 0.20659937523305416,
      "rewards/cosine_scaled_reward": -0.10727175627835095,
      "rewards/format_reward": 0.5000000074505806,
      "step": 53
    },
    {
      "completion_length": 3126.1146240234375,
      "epoch": 0.09254498714652956,
      "grad_norm": 0.17827318608760834,
      "kl": 0.001310110092163086,
      "learning_rate": 5.657047735161255e-07,
      "loss": 0.0001,
      "reward": 0.0032570459879934788,
      "reward_std": 0.15636450424790382,
      "rewards/cosine_scaled_reward": -0.14380789827555418,
      "rewards/format_reward": 0.30208333395421505,
      "step": 54
    },
    {
      "completion_length": 2929.9688720703125,
      "epoch": 0.09425878320479864,
      "grad_norm": 0.1380075216293335,
      "kl": 0.001453399658203125,
      "learning_rate": 5.5e-07,
      "loss": 0.0001,
      "reward": 0.045829253271222115,
      "reward_std": 0.1648220308125019,
      "rewards/cosine_scaled_reward": -0.06621671980246902,
      "rewards/format_reward": 0.31250000558793545,
      "step": 55
    },
    {
      "completion_length": 2777.7083435058594,
      "epoch": 0.0959725792630677,
      "grad_norm": 0.15538543462753296,
      "kl": 0.0006000995635986328,
      "learning_rate": 5.342952264838747e-07,
      "loss": 0.0,
      "reward": 0.05388350097928196,
      "reward_std": 0.1790449097752571,
      "rewards/cosine_scaled_reward": -0.09807816706597805,
      "rewards/format_reward": 0.4062500186264515,
      "step": 56
    },
    {
      "completion_length": 3187.697998046875,
      "epoch": 0.09768637532133675,
      "grad_norm": 0.10485029965639114,
      "kl": 0.0006384849548339844,
      "learning_rate": 5.186095868151436e-07,
      "loss": 0.0,
      "reward": 0.02806462941225618,
      "reward_std": 0.1900525726377964,
      "rewards/cosine_scaled_reward": -0.10978525504469872,
      "rewards/format_reward": 0.3333333432674408,
      "step": 57
    },
    {
      "completion_length": 2514.208465576172,
      "epoch": 0.09940017137960583,
      "grad_norm": 0.15051576495170593,
      "kl": 0.0010862350463867188,
      "learning_rate": 5.02962191529556e-07,
      "loss": 0.0,
      "reward": 0.04878209065645933,
      "reward_std": 0.17859612964093685,
      "rewards/cosine_scaled_reward": -0.17480848729610443,
      "rewards/format_reward": 0.5416666865348816,
      "step": 58
    },
    {
      "completion_length": 2698.4896240234375,
      "epoch": 0.10111396743787489,
      "grad_norm": 0.17716623842716217,
      "kl": 0.00115966796875,
      "learning_rate": 4.873721045679706e-07,
      "loss": 0.0,
      "reward": 0.18329142779111862,
      "reward_std": 0.27025362849235535,
      "rewards/cosine_scaled_reward": 0.0847717362921685,
      "rewards/format_reward": 0.5520833507180214,
      "step": 59
    },
    {
      "completion_length": 2798.9375610351562,
      "epoch": 0.10282776349614396,
      "grad_norm": 0.17029182612895966,
      "kl": 0.00177001953125,
      "learning_rate": 4.7185832004988133e-07,
      "loss": 0.0001,
      "reward": 0.02042369358241558,
      "reward_std": 0.13106655701994896,
      "rewards/cosine_scaled_reward": -0.17489198595285416,
      "rewards/format_reward": 0.4270833395421505,
      "step": 60
    },
    {
      "completion_length": 2818.84375,
      "epoch": 0.10454155955441302,
      "grad_norm": 0.14482490718364716,
      "kl": 0.0011817216873168945,
      "learning_rate": 4.5643973913200837e-07,
      "loss": 0.0,
      "reward": 0.06702410569414496,
      "reward_std": 0.16219814494252205,
      "rewards/cosine_scaled_reward": -0.07603749074041843,
      "rewards/format_reward": 0.4166666865348816,
      "step": 61
    },
    {
      "completion_length": 3321.2604370117188,
      "epoch": 0.1062553556126821,
      "grad_norm": 0.12993496656417847,
      "kl": 0.0014019012451171875,
      "learning_rate": 4.4113514698014953e-07,
      "loss": 0.0001,
      "reward": -0.06851966120302677,
      "reward_std": 0.1139833852648735,
      "rewards/cosine_scaled_reward": -0.20644466951489449,
      "rewards/format_reward": 0.14583333861082792,
      "step": 62
    },
    {
      "completion_length": 3189.135498046875,
      "epoch": 0.10796915167095116,
      "grad_norm": 0.1303822249174118,
      "kl": 0.0011129379272460938,
      "learning_rate": 4.2596318988235037e-07,
      "loss": 0.0,
      "reward": -0.018826739862561226,
      "reward_std": 0.13178249262273312,
      "rewards/cosine_scaled_reward": -0.13564759586006403,
      "rewards/format_reward": 0.19791666977107525,
      "step": 63
    },
    {
      "completion_length": 2808.697998046875,
      "epoch": 0.10968294772922023,
      "grad_norm": 0.1368769407272339,
      "kl": 0.0009212493896484375,
      "learning_rate": 4.1094235253127374e-07,
      "loss": 0.0,
      "reward": 0.05554246995598078,
      "reward_std": 0.17365428805351257,
      "rewards/cosine_scaled_reward": -0.0875049065798521,
      "rewards/format_reward": 0.3958333460614085,
      "step": 64
    },
    {
      "completion_length": 3047.6771240234375,
      "epoch": 0.11139674378748929,
      "grad_norm": 0.11966266483068466,
      "kl": 0.0005755424499511719,
      "learning_rate": 3.9609093550344907e-07,
      "loss": 0.0,
      "reward": 0.08035540394484997,
      "reward_std": 0.1835373416543007,
      "rewards/cosine_scaled_reward": -0.03351259673945606,
      "rewards/format_reward": 0.385416679084301,
      "step": 65
    },
    {
      "completion_length": 2938.4166870117188,
      "epoch": 0.11311053984575835,
      "grad_norm": 0.15843810141086578,
      "kl": 0.0007996559143066406,
      "learning_rate": 3.8142703296283953e-07,
      "loss": 0.0,
      "reward": 0.02124944794923067,
      "reward_std": 0.14316299185156822,
      "rewards/cosine_scaled_reward": -0.09920864924788475,
      "rewards/format_reward": 0.28125,
      "step": 66
    },
    {
      "completion_length": 2877.0834350585938,
      "epoch": 0.11482433590402742,
      "grad_norm": 0.16664999723434448,
      "kl": 0.0018649101257324219,
      "learning_rate": 3.6696851061588994e-07,
      "loss": 0.0001,
      "reward": 0.023960473015904427,
      "reward_std": 0.17024591751396656,
      "rewards/cosine_scaled_reward": -0.12505127862095833,
      "rewards/format_reward": 0.3437500102445483,
      "step": 67
    },
    {
      "completion_length": 2517.4479370117188,
      "epoch": 0.11653813196229648,
      "grad_norm": 0.18517287075519562,
      "kl": 0.0015734434127807617,
      "learning_rate": 3.5273298394491515e-07,
      "loss": 0.0001,
      "reward": 0.12474549748003483,
      "reward_std": 0.17121394164860249,
      "rewards/cosine_scaled_reward": -0.03634676709771156,
      "rewards/format_reward": 0.5625000074505806,
      "step": 68
    },
    {
      "completion_length": 2816.3646850585938,
      "epoch": 0.11825192802056556,
      "grad_norm": 0.16400621831417084,
      "kl": 0.0013251304626464844,
      "learning_rate": 3.387377967463493e-07,
      "loss": 0.0001,
      "reward": 0.009894740185700357,
      "reward_std": 0.17024565115571022,
      "rewards/cosine_scaled_reward": -0.16932371503207833,
      "rewards/format_reward": 0.375,
      "step": 69
    },
    {
      "completion_length": 2789.4168090820312,
      "epoch": 0.11996572407883462,
      "grad_norm": 0.1394311487674713,
      "kl": 0.0010738372802734375,
      "learning_rate": 3.250000000000001e-07,
      "loss": 0.0,
      "reward": 0.1316700098104775,
      "reward_std": 0.2457997277379036,
      "rewards/cosine_scaled_reward": 0.04584646038711071,
      "rewards/format_reward": 0.4270833395421505,
      "step": 70
    },
    {
      "completion_length": 2267.166717529297,
      "epoch": 0.12167952013710369,
      "grad_norm": 0.15189269185066223,
      "kl": 0.001094818115234375,
      "learning_rate": 3.115363310950578e-07,
      "loss": 0.0,
      "reward": 0.24627735652029514,
      "reward_std": 0.1724311113357544,
      "rewards/cosine_scaled_reward": 0.16811909526586533,
      "rewards/format_reward": 0.6354166865348816,
      "step": 71
    },
    {
      "completion_length": 3065.7501220703125,
      "epoch": 0.12339331619537275,
      "grad_norm": 0.1408892422914505,
      "kl": 0.0006256103515625,
      "learning_rate": 2.9836319343816397e-07,
      "loss": 0.0,
      "reward": 0.0009416388347744942,
      "reward_std": 0.1657393015921116,
      "rewards/cosine_scaled_reward": -0.15875469404272735,
      "rewards/format_reward": 0.3229166744276881,
      "step": 72
    },
    {
      "completion_length": 2957.0625610351562,
      "epoch": 0.12510711225364182,
      "grad_norm": 0.13583190739154816,
      "kl": 0.00033473968505859375,
      "learning_rate": 2.854966364683872e-07,
      "loss": 0.0,
      "reward": 0.06538257002830505,
      "reward_std": 0.1970806121826172,
      "rewards/cosine_scaled_reward": -0.09460783936083317,
      "rewards/format_reward": 0.447916679084301,
      "step": 73
    },
    {
      "completion_length": 3106.010498046875,
      "epoch": 0.1268209083119109,
      "grad_norm": 0.15821737051010132,
      "kl": 0.0009412765502929688,
      "learning_rate": 2.729523361034538e-07,
      "loss": 0.0,
      "reward": 0.0028651878237724304,
      "reward_std": 0.19838279485702515,
      "rewards/cosine_scaled_reward": -0.13493257109075785,
      "rewards/format_reward": 0.2812500111758709,
      "step": 74
    },
    {
      "completion_length": 2847.541748046875,
      "epoch": 0.12853470437017994,
      "grad_norm": 0.16942313313484192,
      "kl": 0.0019121170043945312,
      "learning_rate": 2.6074557564105724e-07,
      "loss": 0.0001,
      "reward": 0.0911966897547245,
      "reward_std": 0.21281348168849945,
      "rewards/cosine_scaled_reward": -0.028547534719109535,
      "rewards/format_reward": 0.41666668467223644,
      "step": 75
    },
    {
      "completion_length": 2846.9896850585938,
      "epoch": 0.13024850042844902,
      "grad_norm": 0.1670897901058197,
      "kl": 0.0010271072387695312,
      "learning_rate": 2.488912271385139e-07,
      "loss": 0.0,
      "reward": 0.02618580963462591,
      "reward_std": 0.15880979225039482,
      "rewards/cosine_scaled_reward": -0.18800780980382115,
      "rewards/format_reward": 0.4791666828095913,
      "step": 76
    },
    {
      "completion_length": 3011.2188110351562,
      "epoch": 0.1319622964867181,
      "grad_norm": 0.1846207082271576,
      "kl": 0.0016541481018066406,
      "learning_rate": 2.374037332934512e-07,
      "loss": 0.0001,
      "reward": 0.01657121942844242,
      "reward_std": 0.19087375327944756,
      "rewards/cosine_scaled_reward": -0.10442618932574987,
      "rewards/format_reward": 0.27083334513008595,
      "step": 77
    },
    {
      "completion_length": 3298.625,
      "epoch": 0.13367609254498714,
      "grad_norm": 0.12679187953472137,
      "kl": 0.0008182525634765625,
      "learning_rate": 2.2629708984760706e-07,
      "loss": 0.0,
      "reward": -0.07863575895316899,
      "reward_std": 0.11674198694527149,
      "rewards/cosine_scaled_reward": -0.24837390705943108,
      "rewards/format_reward": 0.1875000074505806,
      "step": 78
    },
    {
      "completion_length": 3014.0521850585938,
      "epoch": 0.1353898886032562,
      "grad_norm": 0.19050873816013336,
      "kl": 0.0012989044189453125,
      "learning_rate": 2.1558482853517253e-07,
      "loss": 0.0001,
      "reward": 0.14426555391401052,
      "reward_std": 0.21018010936677456,
      "rewards/cosine_scaled_reward": 0.08679820597171783,
      "rewards/format_reward": 0.3958333507180214,
      "step": 79
    },
    {
      "completion_length": 2568.697998046875,
      "epoch": 0.13710368466152528,
      "grad_norm": 0.14511092007160187,
      "kl": 0.0016317367553710938,
      "learning_rate": 2.0528000059645995e-07,
      "loss": 0.0001,
      "reward": 0.15173329645767808,
      "reward_std": 0.1817643828690052,
      "rewards/cosine_scaled_reward": 0.00575945433229208,
      "rewards/format_reward": 0.5833333432674408,
      "step": 80
    },
    {
      "completion_length": 2273.5209045410156,
      "epoch": 0.13881748071979436,
      "grad_norm": 0.1482679694890976,
      "kl": 0.0021762847900390625,
      "learning_rate": 1.9539516087697517e-07,
      "loss": 0.0001,
      "reward": 0.17546646296977997,
      "reward_std": 0.18799066171050072,
      "rewards/cosine_scaled_reward": 0.05814665276557207,
      "rewards/format_reward": 0.5729166865348816,
      "step": 81
    },
    {
      "completion_length": 2761.7709350585938,
      "epoch": 0.1405312767780634,
      "grad_norm": 0.14379660785198212,
      "kl": 0.000904083251953125,
      "learning_rate": 1.8594235253127372e-07,
      "loss": 0.0,
      "reward": 0.09056159481406212,
      "reward_std": 0.19218357279896736,
      "rewards/cosine_scaled_reward": -0.07696967083029449,
      "rewards/format_reward": 0.5104166865348816,
      "step": 82
    },
    {
      "completion_length": 2424.1666870117188,
      "epoch": 0.14224507283633248,
      "grad_norm": 0.18656377494335175,
      "kl": 0.0023345947265625,
      "learning_rate": 1.7693309235023127e-07,
      "loss": 0.0001,
      "reward": 0.17556490686547477,
      "reward_std": 0.24106226861476898,
      "rewards/cosine_scaled_reward": 0.06774241663515568,
      "rewards/format_reward": 0.5520833432674408,
      "step": 83
    },
    {
      "completion_length": 2954.041748046875,
      "epoch": 0.14395886889460155,
      "grad_norm": 0.12463019788265228,
      "kl": 0.0012934207916259766,
      "learning_rate": 1.6837835672960831e-07,
      "loss": 0.0001,
      "reward": 0.07739502191543579,
      "reward_std": 0.14365380816161633,
      "rewards/cosine_scaled_reward": -0.024237127974629402,
      "rewards/format_reward": 0.3541666716337204,
      "step": 84
    },
    {
      "completion_length": 3007.1771850585938,
      "epoch": 0.1456726649528706,
      "grad_norm": 0.1296854168176651,
      "kl": 0.0006504058837890625,
      "learning_rate": 1.6028856829700258e-07,
      "loss": 0.0,
      "reward": 0.06423507491126657,
      "reward_std": 0.18408943712711334,
      "rewards/cosine_scaled_reward": -0.08204184658825397,
      "rewards/format_reward": 0.416666679084301,
      "step": 85
    },
    {
      "completion_length": 3371.8646240234375,
      "epoch": 0.14738646101113967,
      "grad_norm": 0.1433609277009964,
      "kl": 0.0012350082397460938,
      "learning_rate": 1.5267358321348285e-07,
      "loss": 0.0,
      "reward": -0.03188338316977024,
      "reward_std": 0.16921599954366684,
      "rewards/cosine_scaled_reward": -0.13588787615299225,
      "rewards/format_reward": 0.14583333861082792,
      "step": 86
    },
    {
      "completion_length": 3052.2501220703125,
      "epoch": 0.14910025706940874,
      "grad_norm": 0.13880982995033264,
      "kl": 0.001392364501953125,
      "learning_rate": 1.4554267916537495e-07,
      "loss": 0.0001,
      "reward": -0.018908033147454262,
      "reward_std": 0.1546711027622223,
      "rewards/cosine_scaled_reward": -0.18234001845121384,
      "rewards/format_reward": 0.2916666753590107,
      "step": 87
    },
    {
      "completion_length": 2693.0626220703125,
      "epoch": 0.15081405312767782,
      "grad_norm": 0.15144215524196625,
      "kl": 0.0007927417755126953,
      "learning_rate": 1.3890454406082956e-07,
      "loss": 0.0,
      "reward": 0.12248719483613968,
      "reward_std": 0.18367315083742142,
      "rewards/cosine_scaled_reward": 0.043038370087742805,
      "rewards/format_reward": 0.3958333507180214,
      "step": 88
    },
    {
      "completion_length": 3063.604248046875,
      "epoch": 0.15252784918594686,
      "grad_norm": 0.14191189408302307,
      "kl": 0.0008802413940429688,
      "learning_rate": 1.3276726544494571e-07,
      "loss": 0.0,
      "reward": 0.0365169154247269,
      "reward_std": 0.15275101736187935,
      "rewards/cosine_scaled_reward": -0.0832907222211361,
      "rewards/format_reward": 0.31250000558793545,
      "step": 89
    },
    {
      "completion_length": 2311.010467529297,
      "epoch": 0.15424164524421594,
      "grad_norm": 0.17067813873291016,
      "kl": 0.0018815994262695312,
      "learning_rate": 1.2713832064634125e-07,
      "loss": 0.0001,
      "reward": 0.1953146681189537,
      "reward_std": 0.2011973224580288,
      "rewards/cosine_scaled_reward": 0.08297107368707657,
      "rewards/format_reward": 0.6041666716337204,
      "step": 90
    },
    {
      "completion_length": 2986.729248046875,
      "epoch": 0.155955441302485,
      "grad_norm": 0.1307849884033203,
      "kl": 0.0007395744323730469,
      "learning_rate": 1.220245676671809e-07,
      "loss": 0.0,
      "reward": 0.05273827910423279,
      "reward_std": 0.22075917199254036,
      "rewards/cosine_scaled_reward": -0.054260533303022385,
      "rewards/format_reward": 0.3125000102445483,
      "step": 91
    },
    {
      "completion_length": 2813.3543090820312,
      "epoch": 0.15766923736075408,
      "grad_norm": 0.13788200914859772,
      "kl": 0.0018596649169921875,
      "learning_rate": 1.1743223682775649e-07,
      "loss": 0.0001,
      "reward": 0.02730799838900566,
      "reward_std": 0.15018530935049057,
      "rewards/cosine_scaled_reward": -0.13331225764704868,
      "rewards/format_reward": 0.375,
      "step": 92
    },
    {
      "completion_length": 3271.2500610351562,
      "epoch": 0.15938303341902313,
      "grad_norm": 0.1422564685344696,
      "kl": 0.001224517822265625,
      "learning_rate": 1.1336692317580158e-07,
      "loss": 0.0,
      "reward": 0.08444089069962502,
      "reward_std": 0.18314170092344284,
      "rewards/cosine_scaled_reward": 0.026102278381586075,
      "rewards/format_reward": 0.2812500037252903,
      "step": 93
    },
    {
      "completion_length": 3040.9688720703125,
      "epoch": 0.1610968294772922,
      "grad_norm": 0.1662042886018753,
      "kl": 0.0022430419921875,
      "learning_rate": 1.0983357966978745e-07,
      "loss": 0.0001,
      "reward": 0.028276806697249413,
      "reward_std": 0.17833666875958443,
      "rewards/cosine_scaled_reward": -0.14247922599315643,
      "rewards/format_reward": 0.3958333544433117,
      "step": 94
    },
    {
      "completion_length": 3048.197998046875,
      "epoch": 0.16281062553556128,
      "grad_norm": 0.12611675262451172,
      "kl": 0.0008878707885742188,
      "learning_rate": 1.068365111445064e-07,
      "loss": 0.0,
      "reward": 0.07369702425785363,
      "reward_std": 0.18385101296007633,
      "rewards/cosine_scaled_reward": -0.07403519004583359,
      "rewards/format_reward": 0.4375000074505806,
      "step": 95
    },
    {
      "completion_length": 3014.7918090820312,
      "epoch": 0.16452442159383032,
      "grad_norm": 0.16144989430904388,
      "kl": 0.0016870498657226562,
      "learning_rate": 1.0437936906629334e-07,
      "loss": 0.0001,
      "reward": 0.07583070383407176,
      "reward_std": 0.21927380189299583,
      "rewards/cosine_scaled_reward": -0.03254310414195061,
      "rewards/format_reward": 0.3645833395421505,
      "step": 96
    },
    {
      "completion_length": 2385.0521240234375,
      "epoch": 0.1662382176520994,
      "grad_norm": 0.15538226068019867,
      "kl": 0.0020532608032226562,
      "learning_rate": 1.0246514708427701e-07,
      "loss": 0.0001,
      "reward": 0.09558990353252739,
      "reward_std": 0.14071721211075783,
      "rewards/cosine_scaled_reward": -0.1022979412227869,
      "rewards/format_reward": 0.5833333507180214,
      "step": 97
    },
    {
      "completion_length": 2722.2500610351562,
      "epoch": 0.16795201371036847,
      "grad_norm": 0.16014371812343597,
      "kl": 0.0013885498046875,
      "learning_rate": 1.0109617738307911e-07,
      "loss": 0.0001,
      "reward": 0.024186712922528386,
      "reward_std": 0.16764022037386894,
      "rewards/cosine_scaled_reward": -0.1459345854818821,
      "rewards/format_reward": 0.385416679084301,
      "step": 98
    },
    {
      "completion_length": 2599.0625610351562,
      "epoch": 0.16966580976863754,
      "grad_norm": 0.132161945104599,
      "kl": 0.0023794174194335938,
      "learning_rate": 1.002741278414069e-07,
      "loss": 0.0001,
      "reward": 0.11149843037128448,
      "reward_std": 0.1786147654056549,
      "rewards/cosine_scaled_reward": -0.06053047860041261,
      "rewards/format_reward": 0.5625000149011612,
      "step": 99
    },
    {
      "completion_length": 2782.8126220703125,
      "epoch": 0.1713796058269066,
      "grad_norm": 0.15893962979316711,
      "kl": 0.0020580291748046875,
      "learning_rate": 1e-07,
      "loss": 0.0001,
      "reward": 0.0819746982306242,
      "reward_std": 0.2312169447541237,
      "rewards/cosine_scaled_reward": -0.04185225034598261,
      "rewards/format_reward": 0.4062500074505806,
      "step": 100
    },
    {
      "epoch": 0.1713796058269066,
      "step": 100,
      "total_flos": 0.0,
      "train_loss": 3.568841239768084e-05,
      "train_runtime": 22061.4799,
      "train_samples_per_second": 0.435,
      "train_steps_per_second": 0.005
    }
  ],
  "logging_steps": 1,
  "max_steps": 100,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 50,
  "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": 4,
  "trial_name": null,
  "trial_params": null
}