data/log_history_run3.json

[
  {
    "loss": 0.0,
    "grad_norm": 0.0,
    "learning_rate": 5e-06,
    "num_tokens": 3013.0,
    "completions/mean_length": 137.25,
    "completions/min_length": 109.0,
    "completions/max_length": 182.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 137.25,
    "completions/min_terminated_length": 109.0,
    "completions/max_terminated_length": 182.0,
    "rewards/reward_fn/mean": 0.33500000834465027,
    "rewards/reward_fn/std": 0.0,
    "reward": 0.33500000834465027,
    "reward_std": 0.0,
    "frac_reward_zero_std": 1.0,
    "entropy": 1.2721036672592163,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 26.99397590800072,
    "epoch": 0.0016666666666666668,
    "step": 1
  },
  {
    "loss": -0.04769711568951607,
    "grad_norm": 0.376953125,
    "learning_rate": 4.975000000000001e-06,
    "num_tokens": 5917.0,
    "completions/mean_length": 114.0,
    "completions/min_length": 98.0,
    "completions/max_length": 125.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 114.0,
    "completions/min_terminated_length": 98.0,
    "completions/max_terminated_length": 125.0,
    "rewards/reward_fn/mean": 0.9693750143051147,
    "rewards/reward_fn/std": 0.00875002145767212,
    "reward": 0.9693750143051147,
    "reward_std": 0.00875002145767212,
    "frac_reward_zero_std": 0.0,
    "entropy": 0.901360273361206,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 19.053884660999756,
    "epoch": 0.0033333333333333335,
    "step": 2
  },
  {
    "loss": -0.011300479993224144,
    "grad_norm": 0.28515625,
    "learning_rate": 4.95e-06,
    "num_tokens": 8552.0,
    "completions/mean_length": 143.75,
    "completions/min_length": 114.0,
    "completions/max_length": 197.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 143.75,
    "completions/min_terminated_length": 114.0,
    "completions/max_terminated_length": 197.0,
    "rewards/reward_fn/mean": 0.4793750047683716,
    "rewards/reward_fn/std": 0.288750022649765,
    "reward": 0.4793750047683716,
    "reward_std": 0.28874996304512024,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.1808185577392578,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 28.53829234299701,
    "epoch": 0.005,
    "step": 3
  },
  {
    "loss": 0.07140593230724335,
    "grad_norm": 0.32421875,
    "learning_rate": 4.925e-06,
    "num_tokens": 11250.0,
    "completions/mean_length": 115.5,
    "completions/min_length": 99.0,
    "completions/max_length": 132.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 115.5,
    "completions/min_terminated_length": 99.0,
    "completions/max_terminated_length": 132.0,
    "rewards/reward_fn/mean": 0.4925000071525574,
    "rewards/reward_fn/std": 0.3149999678134918,
    "reward": 0.4925000071525574,
    "reward_std": 0.3149999976158142,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0730382204055786,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 20.0260389340001,
    "epoch": 0.006666666666666667,
    "step": 4
  },
  {
    "loss": 0.005974027793854475,
    "grad_norm": 0.30078125,
    "learning_rate": 4.9000000000000005e-06,
    "num_tokens": 14256.0,
    "completions/mean_length": 125.5,
    "completions/min_length": 106.0,
    "completions/max_length": 151.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 125.5,
    "completions/min_terminated_length": 106.0,
    "completions/max_terminated_length": 151.0,
    "rewards/reward_fn/mean": 0.48374998569488525,
    "rewards/reward_fn/std": 0.29750001430511475,
    "reward": 0.48374998569488525,
    "reward_std": 0.29750001430511475,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.044655680656433,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 22.637786576000508,
    "epoch": 0.008333333333333333,
    "step": 5
  },
  {
    "loss": -0.018765343353152275,
    "grad_norm": 0.30859375,
    "learning_rate": 4.875e-06,
    "num_tokens": 17146.0,
    "completions/mean_length": 146.5,
    "completions/min_length": 120.0,
    "completions/max_length": 175.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 146.5,
    "completions/min_terminated_length": 120.0,
    "completions/max_terminated_length": 175.0,
    "rewards/reward_fn/mean": 0.4925000071525574,
    "rewards/reward_fn/std": 0.3149999678134918,
    "reward": 0.4925000071525574,
    "reward_std": 0.3149999976158142,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.157387137413025,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 25.646922176998487,
    "epoch": 0.01,
    "step": 6
  },
  {
    "loss": 0.04119409993290901,
    "grad_norm": 0.28515625,
    "learning_rate": 4.85e-06,
    "num_tokens": 20239.0,
    "completions/mean_length": 167.25,
    "completions/min_length": 135.0,
    "completions/max_length": 185.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 167.25,
    "completions/min_terminated_length": 135.0,
    "completions/max_terminated_length": 185.0,
    "rewards/reward_fn/mean": 0.6412500143051147,
    "rewards/reward_fn/std": 0.35391557216644287,
    "reward": 0.6412500143051147,
    "reward_std": 0.35391560196876526,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.1936835050582886,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 27.48416990199985,
    "epoch": 0.011666666666666667,
    "step": 7
  },
  {
    "loss": 0.08870118111371994,
    "grad_norm": 0.23046875,
    "learning_rate": 4.825e-06,
    "num_tokens": 22611.0,
    "completions/mean_length": 168.0,
    "completions/min_length": 136.0,
    "completions/max_length": 221.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 168.0,
    "completions/min_terminated_length": 136.0,
    "completions/max_terminated_length": 221.0,
    "rewards/reward_fn/mean": 0.500374972820282,
    "rewards/reward_fn/std": 0.31010571122169495,
    "reward": 0.500374972820282,
    "reward_std": 0.31010571122169495,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0776662826538086,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 31.74235246600074,
    "epoch": 0.013333333333333334,
    "step": 8
  },
  {
    "loss": 0.0,
    "grad_norm": 0.0,
    "learning_rate": 4.800000000000001e-06,
    "num_tokens": 25760.0,
    "completions/mean_length": 159.25,
    "completions/min_length": 107.0,
    "completions/max_length": 212.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 159.25,
    "completions/min_terminated_length": 107.0,
    "completions/max_terminated_length": 212.0,
    "rewards/reward_fn/mean": 0.33500000834465027,
    "rewards/reward_fn/std": 0.0,
    "reward": 0.33500000834465027,
    "reward_std": 0.0,
    "frac_reward_zero_std": 1.0,
    "entropy": 1.1662333011627197,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 31.728966909002338,
    "epoch": 0.015,
    "step": 9
  },
  {
    "loss": 0.1400359570980072,
    "grad_norm": 0.306640625,
    "learning_rate": 4.775e-06,
    "num_tokens": 28494.0,
    "completions/mean_length": 121.5,
    "completions/min_length": 100.0,
    "completions/max_length": 152.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 121.5,
    "completions/min_terminated_length": 100.0,
    "completions/max_terminated_length": 152.0,
    "rewards/reward_fn/mean": 0.7856249809265137,
    "rewards/reward_fn/std": 0.3052483797073364,
    "reward": 0.7856249809265137,
    "reward_std": 0.3052483797073364,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0164639949798584,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 23.630102548999275,
    "epoch": 0.016666666666666666,
    "step": 10
  },
  {
    "loss": 0.04951930046081543,
    "grad_norm": 0.306640625,
    "learning_rate": 4.75e-06,
    "num_tokens": 31562.0,
    "completions/mean_length": 156.0,
    "completions/min_length": 127.0,
    "completions/max_length": 183.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 156.0,
    "completions/min_terminated_length": 127.0,
    "completions/max_terminated_length": 183.0,
    "rewards/reward_fn/mean": 0.8075000047683716,
    "rewards/reward_fn/std": 0.2921329438686371,
    "reward": 0.8075000047683716,
    "reward_std": 0.2921329736709595,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.2048428058624268,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 28.393185649998486,
    "epoch": 0.018333333333333333,
    "step": 11
  },
  {
    "loss": -0.14611932635307312,
    "grad_norm": 0.27734375,
    "learning_rate": 4.7250000000000005e-06,
    "num_tokens": 34288.0,
    "completions/mean_length": 116.5,
    "completions/min_length": 82.0,
    "completions/max_length": 163.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 116.5,
    "completions/min_terminated_length": 82.0,
    "completions/max_terminated_length": 163.0,
    "rewards/reward_fn/mean": 0.33125001192092896,
    "rewards/reward_fn/std": 0.007500007748603821,
    "reward": 0.33125001192092896,
    "reward_std": 0.007500012870877981,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0292104482650757,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 25.381137749998743,
    "epoch": 0.02,
    "step": 12
  },
  {
    "loss": 0.04130696505308151,
    "grad_norm": 0.306640625,
    "learning_rate": 4.7e-06,
    "num_tokens": 37036.0,
    "completions/mean_length": 146.0,
    "completions/min_length": 120.0,
    "completions/max_length": 157.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 146.0,
    "completions/min_terminated_length": 120.0,
    "completions/max_terminated_length": 157.0,
    "rewards/reward_fn/mean": 0.6412500143051147,
    "rewards/reward_fn/std": 0.35391557216644287,
    "reward": 0.6412500143051147,
    "reward_std": 0.35391560196876526,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.2967694997787476,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 24.01959514600094,
    "epoch": 0.021666666666666667,
    "step": 13
  },
  {
    "loss": 0.03943007439374924,
    "grad_norm": 0.279296875,
    "learning_rate": 4.675000000000001e-06,
    "num_tokens": 40060.0,
    "completions/mean_length": 163.0,
    "completions/min_length": 150.0,
    "completions/max_length": 181.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 163.0,
    "completions/min_terminated_length": 150.0,
    "completions/max_terminated_length": 181.0,
    "rewards/reward_fn/mean": 0.9693750143051147,
    "rewards/reward_fn/std": 0.00875002145767212,
    "reward": 0.9693750143051147,
    "reward_std": 0.00875002145767212,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.2760556936264038,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 26.05132091699852,
    "epoch": 0.023333333333333334,
    "step": 14
  },
  {
    "loss": -0.024401623755693436,
    "grad_norm": 0.291015625,
    "learning_rate": 4.65e-06,
    "num_tokens": 42856.0,
    "completions/mean_length": 154.0,
    "completions/min_length": 145.0,
    "completions/max_length": 163.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 154.0,
    "completions/min_terminated_length": 145.0,
    "completions/max_terminated_length": 163.0,
    "rewards/reward_fn/mean": 0.9606249928474426,
    "rewards/reward_fn/std": 0.02202034927904606,
    "reward": 0.9606249928474426,
    "reward_std": 0.02202034927904606,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0788471698760986,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 24.084032504997595,
    "epoch": 0.025,
    "step": 15
  },
  {
    "loss": 0.10702173411846161,
    "grad_norm": 0.275390625,
    "learning_rate": 4.625000000000001e-06,
    "num_tokens": 46438.0,
    "completions/mean_length": 193.5,
    "completions/min_length": 159.0,
    "completions/max_length": 257.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 193.5,
    "completions/min_terminated_length": 159.0,
    "completions/max_terminated_length": 257.0,
    "rewards/reward_fn/mean": 0.6493749618530273,
    "rewards/reward_fn/std": 0.35504916310310364,
    "reward": 0.6493749618530273,
    "reward_std": 0.355049192905426,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.4748042821884155,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 37.7986828969988,
    "epoch": 0.02666666666666667,
    "step": 16
  },
  {
    "loss": 0.04349641874432564,
    "grad_norm": 0.330078125,
    "learning_rate": 4.600000000000001e-06,
    "num_tokens": 48760.0,
    "completions/mean_length": 143.5,
    "completions/min_length": 117.0,
    "completions/max_length": 156.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 143.5,
    "completions/min_terminated_length": 117.0,
    "completions/max_terminated_length": 156.0,
    "rewards/reward_fn/mean": 0.9275000095367432,
    "rewards/reward_fn/std": 0.07499998807907104,
    "reward": 0.9275000095367432,
    "reward_std": 0.07499998807907104,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.177292823791504,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 22.96707353400052,
    "epoch": 0.028333333333333332,
    "step": 17
  },
  {
    "loss": -0.0374990850687027,
    "grad_norm": 0.4375,
    "learning_rate": 4.575e-06,
    "num_tokens": 51631.0,
    "completions/mean_length": 108.75,
    "completions/min_length": 93.0,
    "completions/max_length": 122.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 108.75,
    "completions/min_terminated_length": 93.0,
    "completions/max_terminated_length": 122.0,
    "rewards/reward_fn/mean": 0.9693750143051147,
    "rewards/reward_fn/std": 0.00875002145767212,
    "reward": 0.9693750143051147,
    "reward_std": 0.00875002145767212,
    "frac_reward_zero_std": 0.0,
    "entropy": 0.9841785430908203,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 18.896498049998627,
    "epoch": 0.03,
    "step": 18
  },
  {
    "loss": -0.02571682445704937,
    "grad_norm": 0.375,
    "learning_rate": 4.5500000000000005e-06,
    "num_tokens": 54431.0,
    "completions/mean_length": 116.0,
    "completions/min_length": 107.0,
    "completions/max_length": 125.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 116.0,
    "completions/min_terminated_length": 107.0,
    "completions/max_terminated_length": 125.0,
    "rewards/reward_fn/mean": 0.9562499523162842,
    "rewards/reward_fn/std": 0.017499983310699463,
    "reward": 0.9562499523162842,
    "reward_std": 0.017499983310699463,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.1009763479232788,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 19.452795449000405,
    "epoch": 0.03166666666666667,
    "step": 19
  },
  {
    "loss": 0.1016233041882515,
    "grad_norm": 0.361328125,
    "learning_rate": 4.525000000000001e-06,
    "num_tokens": 57738.0,
    "completions/mean_length": 140.75,
    "completions/min_length": 85.0,
    "completions/max_length": 187.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 140.75,
    "completions/min_terminated_length": 85.0,
    "completions/max_terminated_length": 187.0,
    "rewards/reward_fn/mean": 0.953249990940094,
    "rewards/reward_fn/std": 0.01649998500943184,
    "reward": 0.953249990940094,
    "reward_std": 0.01649998500943184,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.152324914932251,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 28.531233167999744,
    "epoch": 0.03333333333333333,
    "step": 20
  },
  {
    "loss": 0.05338107421994209,
    "grad_norm": 0.39453125,
    "learning_rate": 4.5e-06,
    "num_tokens": 60750.0,
    "completions/mean_length": 121.0,
    "completions/min_length": 100.0,
    "completions/max_length": 144.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 121.0,
    "completions/min_terminated_length": 100.0,
    "completions/max_terminated_length": 144.0,
    "rewards/reward_fn/mean": 0.9728749990463257,
    "rewards/reward_fn/std": 0.015750020742416382,
    "reward": 0.9728749990463257,
    "reward_std": 0.015750011429190636,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0693097114562988,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 22.120251809003094,
    "epoch": 0.035,
    "step": 21
  },
  {
    "loss": -0.032417263835668564,
    "grad_norm": 0.37890625,
    "learning_rate": 4.475e-06,
    "num_tokens": 63754.0,
    "completions/mean_length": 122.0,
    "completions/min_length": 106.0,
    "completions/max_length": 146.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 122.0,
    "completions/min_terminated_length": 106.0,
    "completions/max_terminated_length": 146.0,
    "rewards/reward_fn/mean": 0.3306249976158142,
    "rewards/reward_fn/std": 0.008750006556510925,
    "reward": 0.3306249976158142,
    "reward_std": 0.008750001899898052,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0988726615905762,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 22.679130085998622,
    "epoch": 0.03666666666666667,
    "step": 22
  },
  {
    "loss": -0.07120607793331146,
    "grad_norm": 0.306640625,
    "learning_rate": 4.450000000000001e-06,
    "num_tokens": 66768.0,
    "completions/mean_length": 121.5,
    "completions/min_length": 91.0,
    "completions/max_length": 153.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 121.5,
    "completions/min_terminated_length": 91.0,
    "completions/max_terminated_length": 153.0,
    "rewards/reward_fn/mean": 0.9781249761581421,
    "rewards/reward_fn/std": 0.00875002145767212,
    "reward": 0.9781249761581421,
    "reward_std": 0.00875002145767212,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.1439276933670044,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 23.13522673600164,
    "epoch": 0.03833333333333333,
    "step": 23
  },
  {
    "loss": 0.04352416470646858,
    "grad_norm": 0.38671875,
    "learning_rate": 4.425e-06,
    "num_tokens": 69557.0,
    "completions/mean_length": 120.25,
    "completions/min_length": 102.0,
    "completions/max_length": 136.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 120.25,
    "completions/min_terminated_length": 102.0,
    "completions/max_terminated_length": 136.0,
    "rewards/reward_fn/mean": 0.7987499833106995,
    "rewards/reward_fn/std": 0.3096065819263458,
    "reward": 0.7987499833106995,
    "reward_std": 0.3096066117286682,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.2446463108062744,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 20.66519093700117,
    "epoch": 0.04,
    "step": 24
  },
  {
    "loss": -0.06304893642663956,
    "grad_norm": 0.3671875,
    "learning_rate": 4.4e-06,
    "num_tokens": 72543.0,
    "completions/mean_length": 139.5,
    "completions/min_length": 111.0,
    "completions/max_length": 153.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 139.5,
    "completions/min_terminated_length": 111.0,
    "completions/max_terminated_length": 153.0,
    "rewards/reward_fn/mean": 0.6499999761581421,
    "rewards/reward_fn/std": 0.3441172242164612,
    "reward": 0.6499999761581421,
    "reward_std": 0.3441172242164612,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.232834815979004,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 22.765788336000696,
    "epoch": 0.041666666666666664,
    "step": 25
  },
  {
    "loss": 0.0,
    "grad_norm": 0.0,
    "learning_rate": 4.3750000000000005e-06,
    "num_tokens": 74829.0,
    "completions/mean_length": 138.5,
    "completions/min_length": 112.0,
    "completions/max_length": 161.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 138.5,
    "completions/min_terminated_length": 112.0,
    "completions/max_terminated_length": 161.0,
    "rewards/reward_fn/mean": 0.9649999737739563,
    "rewards/reward_fn/std": 0.0,
    "reward": 0.9649999737739563,
    "reward_std": 0.0,
    "frac_reward_zero_std": 1.0,
    "entropy": 1.2499821186065674,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 23.262686510002823,
    "epoch": 0.043333333333333335,
    "step": 26
  },
  {
    "loss": -0.062136210501194,
    "grad_norm": 0.40234375,
    "learning_rate": 4.350000000000001e-06,
    "num_tokens": 78028.0,
    "completions/mean_length": 134.75,
    "completions/min_length": 118.0,
    "completions/max_length": 167.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 134.75,
    "completions/min_terminated_length": 118.0,
    "completions/max_terminated_length": 167.0,
    "rewards/reward_fn/mean": 0.7699999809265137,
    "rewards/reward_fn/std": 0.38999998569488525,
    "reward": 0.7699999809265137,
    "reward_std": 0.38999998569488525,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.2326076030731201,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 25.535409896998317,
    "epoch": 0.045,
    "step": 27
  },
  {
    "loss": 0.10957296937704086,
    "grad_norm": 0.361328125,
    "learning_rate": 4.325e-06,
    "num_tokens": 80670.0,
    "completions/mean_length": 106.5,
    "completions/min_length": 83.0,
    "completions/max_length": 151.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 106.5,
    "completions/min_terminated_length": 83.0,
    "completions/max_terminated_length": 151.0,
    "rewards/reward_fn/mean": 0.9722499847412109,
    "rewards/reward_fn/std": 0.014500021934509277,
    "reward": 0.9722499847412109,
    "reward_std": 0.014500022865831852,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0672565698623657,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 22.534873515000072,
    "epoch": 0.04666666666666667,
    "step": 28
  },
  {
    "loss": -0.1097787395119667,
    "grad_norm": 0.298828125,
    "learning_rate": 4.3e-06,
    "num_tokens": 83466.0,
    "completions/mean_length": 138.0,
    "completions/min_length": 107.0,
    "completions/max_length": 206.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 138.0,
    "completions/min_terminated_length": 107.0,
    "completions/max_terminated_length": 206.0,
    "rewards/reward_fn/mean": 0.8206249475479126,
    "rewards/reward_fn/std": 0.3005298972129822,
    "reward": 0.8206249475479126,
    "reward_std": 0.3005298972129822,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.063458800315857,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 31.418785993999336,
    "epoch": 0.04833333333333333,
    "step": 29
  },
  {
    "loss": 0.020558878779411316,
    "grad_norm": 0.3046875,
    "learning_rate": 4.2750000000000006e-06,
    "num_tokens": 86306.0,
    "completions/mean_length": 146.0,
    "completions/min_length": 127.0,
    "completions/max_length": 170.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 146.0,
    "completions/min_terminated_length": 127.0,
    "completions/max_terminated_length": 170.0,
    "rewards/reward_fn/mean": 0.32625001668930054,
    "rewards/reward_fn/std": 0.010103637352585793,
    "reward": 0.32625001668930054,
    "reward_std": 0.010103637352585793,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.1856608390808105,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 25.58633564399861,
    "epoch": 0.05,
    "step": 30
  },
  {
    "loss": -0.1461215764284134,
    "grad_norm": 0.28515625,
    "learning_rate": 4.25e-06,
    "num_tokens": 89374.0,
    "completions/mean_length": 134.0,
    "completions/min_length": 86.0,
    "completions/max_length": 179.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 134.0,
    "completions/min_terminated_length": 86.0,
    "completions/max_terminated_length": 179.0,
    "rewards/reward_fn/mean": 0.33937501907348633,
    "rewards/reward_fn/std": 0.022020353004336357,
    "reward": 0.33937501907348633,
    "reward_std": 0.022020353004336357,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0142335891723633,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 26.99719125000047,
    "epoch": 0.051666666666666666,
    "step": 31
  },
  {
    "loss": -0.09904350340366364,
    "grad_norm": 0.2734375,
    "learning_rate": 4.225e-06,
    "num_tokens": 92234.0,
    "completions/mean_length": 139.0,
    "completions/min_length": 92.0,
    "completions/max_length": 170.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 139.0,
    "completions/min_terminated_length": 92.0,
    "completions/max_terminated_length": 170.0,
    "rewards/reward_fn/mean": 0.6587499976158142,
    "rewards/reward_fn/std": 0.35391557216644287,
    "reward": 0.6587499976158142,
    "reward_std": 0.35391557216644287,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.3681882619857788,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 25.45712414800073,
    "epoch": 0.05333333333333334,
    "step": 32
  },
  {
    "loss": -0.13209989666938782,
    "grad_norm": 0.359375,
    "learning_rate": 4.2000000000000004e-06,
    "num_tokens": 94781.0,
    "completions/mean_length": 110.75,
    "completions/min_length": 65.0,
    "completions/max_length": 169.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 110.75,
    "completions/min_terminated_length": 65.0,
    "completions/max_terminated_length": 169.0,
    "rewards/reward_fn/mean": 0.37325000762939453,
    "rewards/reward_fn/std": 0.04913502186536789,
    "reward": 0.37325000762939453,
    "reward_std": 0.04913502559065819,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.112715721130371,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 24.924523759000294,
    "epoch": 0.055,
    "step": 33
  },
  {
    "loss": -0.0015746492426842451,
    "grad_norm": 0.298828125,
    "learning_rate": 4.175e-06,
    "num_tokens": 97417.0,
    "completions/mean_length": 169.0,
    "completions/min_length": 149.0,
    "completions/max_length": 196.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 169.0,
    "completions/min_terminated_length": 149.0,
    "completions/max_terminated_length": 196.0,
    "rewards/reward_fn/mean": 0.5913749933242798,
    "rewards/reward_fn/std": 0.31556177139282227,
    "reward": 0.5913749933242798,
    "reward_std": 0.31556180119514465,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.3067529201507568,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 29.42368617000102,
    "epoch": 0.056666666666666664,
    "step": 34
  },
  {
    "loss": 0.0,
    "grad_norm": 0.0,
    "learning_rate": 4.15e-06,
    "num_tokens": 100201.0,
    "completions/mean_length": 128.0,
    "completions/min_length": 93.0,
    "completions/max_length": 152.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 128.0,
    "completions/min_terminated_length": 93.0,
    "completions/max_terminated_length": 152.0,
    "rewards/reward_fn/mean": 0.9649999737739563,
    "rewards/reward_fn/std": 0.0,
    "reward": 0.9649999737739563,
    "reward_std": 0.0,
    "frac_reward_zero_std": 1.0,
    "entropy": 1.064958930015564,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 22.600620116001664,
    "epoch": 0.058333333333333334,
    "step": 35
  },
  {
    "loss": -0.03460807725787163,
    "grad_norm": 0.357421875,
    "learning_rate": 4.125e-06,
    "num_tokens": 102946.0,
    "completions/mean_length": 115.25,
    "completions/min_length": 100.0,
    "completions/max_length": 133.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 115.25,
    "completions/min_terminated_length": 100.0,
    "completions/max_terminated_length": 133.0,
    "rewards/reward_fn/mean": 0.6850000023841858,
    "rewards/reward_fn/std": 0.32457664608955383,
    "reward": 0.6850000023841858,
    "reward_std": 0.32457661628723145,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.1022453308105469,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 19.86391903600088,
    "epoch": 0.06,
    "step": 36
  },
  {
    "loss": 0.0848444476723671,
    "grad_norm": 0.29296875,
    "learning_rate": 4.1e-06,
    "num_tokens": 105380.0,
    "completions/mean_length": 152.5,
    "completions/min_length": 125.0,
    "completions/max_length": 181.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 152.5,
    "completions/min_terminated_length": 125.0,
    "completions/max_terminated_length": 181.0,
    "rewards/reward_fn/mean": 0.6150000095367432,
    "rewards/reward_fn/std": 0.3441172242164612,
    "reward": 0.6150000095367432,
    "reward_std": 0.34411725401878357,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.1503243446350098,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 25.59769222900286,
    "epoch": 0.06166666666666667,
    "step": 37
  },
  {
    "loss": 0.016566012054681778,
    "grad_norm": 0.400390625,
    "learning_rate": 4.075e-06,
    "num_tokens": 108438.0,
    "completions/mean_length": 113.5,
    "completions/min_length": 109.0,
    "completions/max_length": 116.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 113.5,
    "completions/min_terminated_length": 109.0,
    "completions/max_terminated_length": 116.0,
    "rewards/reward_fn/mean": 0.9772499799728394,
    "rewards/reward_fn/std": 0.015256169252097607,
    "reward": 0.9772499799728394,
    "reward_std": 0.015256185084581375,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0994595289230347,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 17.853799561002234,
    "epoch": 0.06333333333333334,
    "step": 38
  },
  {
    "loss": -0.16876700520515442,
    "grad_norm": 0.302734375,
    "learning_rate": 4.05e-06,
    "num_tokens": 111567.0,
    "completions/mean_length": 139.25,
    "completions/min_length": 97.0,
    "completions/max_length": 187.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 139.25,
    "completions/min_terminated_length": 97.0,
    "completions/max_terminated_length": 187.0,
    "rewards/reward_fn/mean": 0.48375001549720764,
    "rewards/reward_fn/std": 0.3209393620491028,
    "reward": 0.48375001549720764,
    "reward_std": 0.3209393620491028,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.2051821947097778,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 27.307055137000134,
    "epoch": 0.065,
    "step": 39
  },
  {
    "loss": 0.09384524822235107,
    "grad_norm": 0.328125,
    "learning_rate": 4.0250000000000004e-06,
    "num_tokens": 114485.0,
    "completions/mean_length": 132.5,
    "completions/min_length": 105.0,
    "completions/max_length": 174.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 132.5,
    "completions/min_terminated_length": 105.0,
    "completions/max_terminated_length": 174.0,
    "rewards/reward_fn/mean": 0.32625001668930054,
    "rewards/reward_fn/std": 0.010103637352585793,
    "reward": 0.32625001668930054,
    "reward_std": 0.010103637352585793,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0574369430541992,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 25.46330049500102,
    "epoch": 0.06666666666666667,
    "step": 40
  },
  {
    "loss": -0.001768031739629805,
    "grad_norm": 0.3359375,
    "learning_rate": 4.000000000000001e-06,
    "num_tokens": 117383.0,
    "completions/mean_length": 127.5,
    "completions/min_length": 119.0,
    "completions/max_length": 138.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 127.5,
    "completions/min_terminated_length": 119.0,
    "completions/max_terminated_length": 138.0,
    "rewards/reward_fn/mean": 0.29750001430511475,
    "rewards/reward_fn/std": 0.09970790892839432,
    "reward": 0.29750001430511475,
    "reward_std": 0.09970790892839432,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.045518159866333,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 20.58559491000051,
    "epoch": 0.06833333333333333,
    "step": 41
  },
  {
    "loss": -0.0037579911295324564,
    "grad_norm": 0.375,
    "learning_rate": 3.975000000000001e-06,
    "num_tokens": 120630.0,
    "completions/mean_length": 136.75,
    "completions/min_length": 110.0,
    "completions/max_length": 156.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 136.75,
    "completions/min_terminated_length": 110.0,
    "completions/max_terminated_length": 156.0,
    "rewards/reward_fn/mean": 0.33937501907348633,
    "rewards/reward_fn/std": 0.036077164113521576,
    "reward": 0.33937501907348633,
    "reward_std": 0.036077164113521576,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.2528449296951294,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 23.60465663499781,
    "epoch": 0.07,
    "step": 42
  },
  {
    "loss": -0.010047557763755322,
    "grad_norm": 0.396484375,
    "learning_rate": 3.95e-06,
    "num_tokens": 123572.0,
    "completions/mean_length": 123.5,
    "completions/min_length": 96.0,
    "completions/max_length": 149.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 123.5,
    "completions/min_terminated_length": 96.0,
    "completions/max_terminated_length": 149.0,
    "rewards/reward_fn/mean": 0.972000002861023,
    "rewards/reward_fn/std": 0.013999998569488525,
    "reward": 0.972000002861023,
    "reward_std": 0.013999998569488525,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0263532400131226,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 21.8416585140003,
    "epoch": 0.07166666666666667,
    "step": 43
  },
  {
    "loss": -0.07597894221544266,
    "grad_norm": 0.294921875,
    "learning_rate": 3.9250000000000005e-06,
    "num_tokens": 126769.0,
    "completions/mean_length": 158.25,
    "completions/min_length": 139.0,
    "completions/max_length": 179.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 158.25,
    "completions/min_terminated_length": 139.0,
    "completions/max_terminated_length": 179.0,
    "rewards/reward_fn/mean": 0.6499999761581421,
    "rewards/reward_fn/std": 0.3441172242164612,
    "reward": 0.6499999761581421,
    "reward_std": 0.3441172242164612,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.4695942401885986,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 26.284113427002012,
    "epoch": 0.07333333333333333,
    "step": 44
  },
  {
    "loss": 0.04559139907360077,
    "grad_norm": 0.37890625,
    "learning_rate": 3.900000000000001e-06,
    "num_tokens": 129512.0,
    "completions/mean_length": 132.75,
    "completions/min_length": 99.0,
    "completions/max_length": 176.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 132.75,
    "completions/min_terminated_length": 99.0,
    "completions/max_terminated_length": 176.0,
    "rewards/reward_fn/mean": 0.6675000190734863,
    "rewards/reward_fn/std": 0.34470999240875244,
    "reward": 0.6675000190734863,
    "reward_std": 0.34470999240875244,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.3445155620574951,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 26.141577066002355,
    "epoch": 0.075,
    "step": 45
  },
  {
    "loss": -0.021484289318323135,
    "grad_norm": 0.349609375,
    "learning_rate": 3.875e-06,
    "num_tokens": 132537.0,
    "completions/mean_length": 125.25,
    "completions/min_length": 115.0,
    "completions/max_length": 131.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 125.25,
    "completions/min_terminated_length": 115.0,
    "completions/max_terminated_length": 131.0,
    "rewards/reward_fn/mean": 0.49162501096725464,
    "rewards/reward_fn/std": 0.3345796763896942,
    "reward": 0.49162501096725464,
    "reward_std": 0.3345796763896942,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.098323941230774,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 20.43934466399878,
    "epoch": 0.07666666666666666,
    "step": 46
  },
  {
    "loss": 0.0,
    "grad_norm": 0.0,
    "learning_rate": 3.85e-06,
    "num_tokens": 135108.0,
    "completions/mean_length": 127.75,
    "completions/min_length": 116.0,
    "completions/max_length": 151.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 127.75,
    "completions/min_terminated_length": 116.0,
    "completions/max_terminated_length": 151.0,
    "rewards/reward_fn/mean": 0.9649999737739563,
    "rewards/reward_fn/std": 0.0,
    "reward": 0.9649999737739563,
    "reward_std": 0.0,
    "frac_reward_zero_std": 1.0,
    "entropy": 1.1513502597808838,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 24.161369523997564,
    "epoch": 0.07833333333333334,
    "step": 47
  },
  {
    "loss": -0.03219185769557953,
    "grad_norm": 0.296875,
    "learning_rate": 3.825000000000001e-06,
    "num_tokens": 137804.0,
    "completions/mean_length": 139.0,
    "completions/min_length": 130.0,
    "completions/max_length": 150.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 139.0,
    "completions/min_terminated_length": 130.0,
    "completions/max_terminated_length": 150.0,
    "rewards/reward_fn/mean": 0.9562499523162842,
    "rewards/reward_fn/std": 0.017499983310699463,
    "reward": 0.9562499523162842,
    "reward_std": 0.017499983310699463,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0865966081619263,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 23.38693464699827,
    "epoch": 0.08,
    "step": 48
  },
  {
    "loss": 0.026122601702809334,
    "grad_norm": 0.35546875,
    "learning_rate": 3.8000000000000005e-06,
    "num_tokens": 140483.0,
    "completions/mean_length": 116.75,
    "completions/min_length": 104.0,
    "completions/max_length": 144.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 116.75,
    "completions/min_terminated_length": 104.0,
    "completions/max_terminated_length": 144.0,
    "rewards/reward_fn/mean": 0.9518749713897705,
    "rewards/reward_fn/std": 0.016754958778619766,
    "reward": 0.9518749713897705,
    "reward_std": 0.01675495319068432,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.214436650276184,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 22.844522692001192,
    "epoch": 0.08166666666666667,
    "step": 49
  },
  {
    "loss": -0.003206648863852024,
    "grad_norm": 0.333984375,
    "learning_rate": 3.7750000000000003e-06,
    "num_tokens": 143371.0,
    "completions/mean_length": 135.0,
    "completions/min_length": 109.0,
    "completions/max_length": 162.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 135.0,
    "completions/min_terminated_length": 109.0,
    "completions/max_terminated_length": 162.0,
    "rewards/reward_fn/mean": 0.6499999761581421,
    "rewards/reward_fn/std": 0.36373066902160645,
    "reward": 0.6499999761581421,
    "reward_std": 0.36373066902160645,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.2262229919433594,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 24.662258178999764,
    "epoch": 0.08333333333333333,
    "step": 50
  },
  {
    "loss": -0.10688165575265884,
    "grad_norm": 0.314453125,
    "learning_rate": 3.7500000000000005e-06,
    "num_tokens": 146097.0,
    "completions/mean_length": 127.5,
    "completions/min_length": 102.0,
    "completions/max_length": 165.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 127.5,
    "completions/min_terminated_length": 102.0,
    "completions/max_terminated_length": 165.0,
    "rewards/reward_fn/mean": 0.9781249761581421,
    "rewards/reward_fn/std": 0.016754990443587303,
    "reward": 0.9781249761581421,
    "reward_std": 0.016754984855651855,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0940418243408203,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 25.447743856999296,
    "epoch": 0.085,
    "step": 51
  },
  {
    "loss": -0.0370216965675354,
    "grad_norm": 0.29296875,
    "learning_rate": 3.7250000000000003e-06,
    "num_tokens": 149049.0,
    "completions/mean_length": 148.0,
    "completions/min_length": 130.0,
    "completions/max_length": 178.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 148.0,
    "completions/min_terminated_length": 130.0,
    "completions/max_terminated_length": 178.0,
    "rewards/reward_fn/mean": 0.9518749713897705,
    "rewards/reward_fn/std": 0.026249976828694344,
    "reward": 0.9518749713897705,
    "reward_std": 0.026249965652823448,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.1700999736785889,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 26.492212680001103,
    "epoch": 0.08666666666666667,
    "step": 52
  },
  {
    "loss": -0.04324590787291527,
    "grad_norm": 0.322265625,
    "learning_rate": 3.7e-06,
    "num_tokens": 151888.0,
    "completions/mean_length": 145.75,
    "completions/min_length": 123.0,
    "completions/max_length": 179.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 145.75,
    "completions/min_terminated_length": 123.0,
    "completions/max_terminated_length": 179.0,
    "rewards/reward_fn/mean": 0.3306249976158142,
    "rewards/reward_fn/std": 0.008750006556510925,
    "reward": 0.3306249976158142,
    "reward_std": 0.008750001899898052,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0578383207321167,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 26.41010283899959,
    "epoch": 0.08833333333333333,
    "step": 53
  },
  {
    "loss": 0.05730345472693443,
    "grad_norm": 0.3125,
    "learning_rate": 3.6750000000000004e-06,
    "num_tokens": 154483.0,
    "completions/mean_length": 131.75,
    "completions/min_length": 117.0,
    "completions/max_length": 147.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 131.75,
    "completions/min_terminated_length": 117.0,
    "completions/max_terminated_length": 147.0,
    "rewards/reward_fn/mean": 0.6499999761581421,
    "rewards/reward_fn/std": 0.3441172242164612,
    "reward": 0.6499999761581421,
    "reward_std": 0.3441172242164612,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.2011436223983765,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 21.763525555998058,
    "epoch": 0.09,
    "step": 54
  },
  {
    "loss": -0.031099891290068626,
    "grad_norm": 0.31640625,
    "learning_rate": 3.65e-06,
    "num_tokens": 157822.0,
    "completions/mean_length": 127.75,
    "completions/min_length": 115.0,
    "completions/max_length": 136.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 127.75,
    "completions/min_terminated_length": 115.0,
    "completions/max_terminated_length": 136.0,
    "rewards/reward_fn/mean": 0.6062500476837158,
    "rewards/reward_fn/std": 0.29886940121650696,
    "reward": 0.6062500476837158,
    "reward_std": 0.29886937141418457,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.1653622388839722,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 21.180196135002916,
    "epoch": 0.09166666666666666,
    "step": 55
  },
  {
    "loss": 0.044428952038288116,
    "grad_norm": 0.400390625,
    "learning_rate": 3.625e-06,
    "num_tokens": 160477.0,
    "completions/mean_length": 110.75,
    "completions/min_length": 94.0,
    "completions/max_length": 125.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 110.75,
    "completions/min_terminated_length": 94.0,
    "completions/max_terminated_length": 125.0,
    "rewards/reward_fn/mean": 0.3193749785423279,
    "rewards/reward_fn/std": 0.01641835644841194,
    "reward": 0.3193749785423279,
    "reward_std": 0.01641835831105709,
    "frac_reward_zero_std": 0.0,
    "entropy": 0.9973193407058716,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 19.055930811002327,
    "epoch": 0.09333333333333334,
    "step": 56
  },
  {
    "loss": 0.0,
    "grad_norm": 0.0,
    "learning_rate": 3.6000000000000003e-06,
    "num_tokens": 163015.0,
    "completions/mean_length": 117.5,
    "completions/min_length": 99.0,
    "completions/max_length": 137.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 117.5,
    "completions/min_terminated_length": 99.0,
    "completions/max_terminated_length": 137.0,
    "rewards/reward_fn/mean": 0.9649999737739563,
    "rewards/reward_fn/std": 0.0,
    "reward": 0.9649999737739563,
    "reward_std": 0.0,
    "frac_reward_zero_std": 1.0,
    "entropy": 1.0432900190353394,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 20.755967883000267,
    "epoch": 0.095,
    "step": 57
  },
  {
    "loss": 0.08609388768672943,
    "grad_norm": 0.392578125,
    "learning_rate": 3.575e-06,
    "num_tokens": 165757.0,
    "completions/mean_length": 127.5,
    "completions/min_length": 108.0,
    "completions/max_length": 149.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 127.5,
    "completions/min_terminated_length": 108.0,
    "completions/max_terminated_length": 149.0,
    "rewards/reward_fn/mean": 0.33937501907348633,
    "rewards/reward_fn/std": 0.022020353004336357,
    "reward": 0.33937501907348633,
    "reward_std": 0.022020353004336357,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0790032148361206,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 22.144303656001284,
    "epoch": 0.09666666666666666,
    "step": 58
  },
  {
    "loss": 0.07765334099531174,
    "grad_norm": 0.314453125,
    "learning_rate": 3.5500000000000003e-06,
    "num_tokens": 168967.0,
    "completions/mean_length": 155.5,
    "completions/min_length": 119.0,
    "completions/max_length": 201.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 155.5,
    "completions/min_terminated_length": 119.0,
    "completions/max_terminated_length": 201.0,
    "rewards/reward_fn/mean": 0.7524999976158142,
    "rewards/reward_fn/std": 0.2626309096813202,
    "reward": 0.7524999976158142,
    "reward_std": 0.2626309096813202,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.3297035694122314,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 29.628764026998397,
    "epoch": 0.09833333333333333,
    "step": 59
  },
  {
    "loss": -0.07361559569835663,
    "grad_norm": 0.24609375,
    "learning_rate": 3.525e-06,
    "num_tokens": 172254.0,
    "completions/mean_length": 181.75,
    "completions/min_length": 130.0,
    "completions/max_length": 262.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 181.75,
    "completions/min_terminated_length": 130.0,
    "completions/max_terminated_length": 262.0,
    "rewards/reward_fn/mean": 0.953374981880188,
    "rewards/reward_fn/std": 0.01649935357272625,
    "reward": 0.953374981880188,
    "reward_std": 0.01649935357272625,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.1880607604980469,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 37.99089504899894,
    "epoch": 0.1,
    "step": 60
  },
  {
    "loss": 0.0,
    "grad_norm": 0.0,
    "learning_rate": 3.5e-06,
    "num_tokens": 175288.0,
    "completions/mean_length": 125.5,
    "completions/min_length": 96.0,
    "completions/max_length": 193.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 125.5,
    "completions/min_terminated_length": 96.0,
    "completions/max_terminated_length": 193.0,
    "rewards/reward_fn/mean": 0.33500000834465027,
    "rewards/reward_fn/std": 0.0,
    "reward": 0.33500000834465027,
    "reward_std": 0.0,
    "frac_reward_zero_std": 1.0,
    "entropy": 1.1564736366271973,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 28.490477613999246,
    "epoch": 0.10166666666666667,
    "step": 61
  },
  {
    "loss": -0.01383125875145197,
    "grad_norm": 0.3359375,
    "learning_rate": 3.475e-06,
    "num_tokens": 177812.0,
    "completions/mean_length": 124.0,
    "completions/min_length": 117.0,
    "completions/max_length": 135.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 124.0,
    "completions/min_terminated_length": 117.0,
    "completions/max_terminated_length": 135.0,
    "rewards/reward_fn/mean": 0.9737499952316284,
    "rewards/reward_fn/std": 0.010103654116392136,
    "reward": 0.9737499952316284,
    "reward_std": 0.010103654116392136,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0834579467773438,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 20.202116815002228,
    "epoch": 0.10333333333333333,
    "step": 62
  },
  {
    "loss": -0.058121208101511,
    "grad_norm": 0.296875,
    "learning_rate": 3.45e-06,
    "num_tokens": 181135.0,
    "completions/mean_length": 156.75,
    "completions/min_length": 132.0,
    "completions/max_length": 186.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 156.75,
    "completions/min_terminated_length": 132.0,
    "completions/max_terminated_length": 186.0,
    "rewards/reward_fn/mean": 0.643875002861023,
    "rewards/reward_fn/std": 0.37093809247016907,
    "reward": 0.643875002861023,
    "reward_std": 0.3709380626678467,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.3051375150680542,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 27.585933769001713,
    "epoch": 0.105,
    "step": 63
  },
  {
    "loss": -0.024057308211922646,
    "grad_norm": 0.359375,
    "learning_rate": 3.4250000000000007e-06,
    "num_tokens": 184026.0,
    "completions/mean_length": 118.75,
    "completions/min_length": 112.0,
    "completions/max_length": 130.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 118.75,
    "completions/min_terminated_length": 112.0,
    "completions/max_terminated_length": 130.0,
    "rewards/reward_fn/mean": 0.32712501287460327,
    "rewards/reward_fn/std": 0.015750005841255188,
    "reward": 0.32712501287460327,
    "reward_std": 0.015750011429190636,
    "frac_reward_zero_std": 0.0,
    "entropy": 0.9585571885108948,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 19.43507832999967,
    "epoch": 0.10666666666666667,
    "step": 64
  },
  {
    "loss": 0.0,
    "grad_norm": 0.0,
    "learning_rate": 3.4000000000000005e-06,
    "num_tokens": 186604.0,
    "completions/mean_length": 124.5,
    "completions/min_length": 112.0,
    "completions/max_length": 147.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 124.5,
    "completions/min_terminated_length": 112.0,
    "completions/max_terminated_length": 147.0,
    "rewards/reward_fn/mean": 0.33500000834465027,
    "rewards/reward_fn/std": 0.0,
    "reward": 0.33500000834465027,
    "reward_std": 0.0,
    "frac_reward_zero_std": 1.0,
    "entropy": 1.098405122756958,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 21.2074764780009,
    "epoch": 0.10833333333333334,
    "step": 65
  },
  {
    "loss": -0.09480708837509155,
    "grad_norm": 0.2890625,
    "learning_rate": 3.3750000000000003e-06,
    "num_tokens": 189773.0,
    "completions/mean_length": 152.25,
    "completions/min_length": 124.0,
    "completions/max_length": 200.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 152.25,
    "completions/min_terminated_length": 124.0,
    "completions/max_terminated_length": 200.0,
    "rewards/reward_fn/mean": 0.3525000214576721,
    "rewards/reward_fn/std": 0.020207257941365242,
    "reward": 0.3525000214576721,
    "reward_std": 0.020207257941365242,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.2371915578842163,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 29.151788433002366,
    "epoch": 0.11,
    "step": 66
  },
  {
    "loss": 0.18502062559127808,
    "grad_norm": 0.341796875,
    "learning_rate": 3.3500000000000005e-06,
    "num_tokens": 192564.0,
    "completions/mean_length": 135.75,
    "completions/min_length": 101.0,
    "completions/max_length": 186.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 135.75,
    "completions/min_terminated_length": 101.0,
    "completions/max_terminated_length": 186.0,
    "rewards/reward_fn/mean": 0.8162499666213989,
    "rewards/reward_fn/std": 0.29749998450279236,
    "reward": 0.8162499666213989,
    "reward_std": 0.29749998450279236,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0945377349853516,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 27.474493719000748,
    "epoch": 0.11166666666666666,
    "step": 67
  },
  {
    "loss": -0.08495711535215378,
    "grad_norm": 0.318359375,
    "learning_rate": 3.3250000000000004e-06,
    "num_tokens": 195718.0,
    "completions/mean_length": 139.5,
    "completions/min_length": 97.0,
    "completions/max_length": 160.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 139.5,
    "completions/min_terminated_length": 97.0,
    "completions/max_terminated_length": 160.0,
    "rewards/reward_fn/mean": 0.6412500143051147,
    "rewards/reward_fn/std": 0.35391557216644287,
    "reward": 0.6412500143051147,
    "reward_std": 0.35391560196876526,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.2198596000671387,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 23.923134037002455,
    "epoch": 0.11333333333333333,
    "step": 68
  },
  {
    "loss": -0.06370344012975693,
    "grad_norm": 0.28515625,
    "learning_rate": 3.3000000000000006e-06,
    "num_tokens": 198864.0,
    "completions/mean_length": 149.5,
    "completions/min_length": 114.0,
    "completions/max_length": 167.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 149.5,
    "completions/min_terminated_length": 114.0,
    "completions/max_terminated_length": 167.0,
    "rewards/reward_fn/mean": 0.6499999761581421,
    "rewards/reward_fn/std": 0.36373066902160645,
    "reward": 0.6499999761581421,
    "reward_std": 0.36373066902160645,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.370664358139038,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 24.89920946699931,
    "epoch": 0.115,
    "step": 69
  },
  {
    "loss": 0.02983768843114376,
    "grad_norm": 0.302734375,
    "learning_rate": 3.2750000000000004e-06,
    "num_tokens": 201847.0,
    "completions/mean_length": 174.75,
    "completions/min_length": 136.0,
    "completions/max_length": 254.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 174.75,
    "completions/min_terminated_length": 136.0,
    "completions/max_terminated_length": 254.0,
    "rewards/reward_fn/mean": 0.48899999260902405,
    "rewards/reward_fn/std": 0.3174019455909729,
    "reward": 0.48899999260902405,
    "reward_std": 0.3174019455909729,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.2808506488800049,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 37.47509274999902,
    "epoch": 0.11666666666666667,
    "step": 70
  },
  {
    "loss": -0.03286806866526604,
    "grad_norm": 0.380859375,
    "learning_rate": 3.2500000000000002e-06,
    "num_tokens": 204520.0,
    "completions/mean_length": 110.25,
    "completions/min_length": 93.0,
    "completions/max_length": 123.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 110.25,
    "completions/min_terminated_length": 93.0,
    "completions/max_terminated_length": 123.0,
    "rewards/reward_fn/mean": 0.824999988079071,
    "rewards/reward_fn/std": 0.2800000011920929,
    "reward": 0.824999988079071,
    "reward_std": 0.2799999713897705,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0120017528533936,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 19.081610291999823,
    "epoch": 0.11833333333333333,
    "step": 71
  },
  {
    "loss": -0.12470149993896484,
    "grad_norm": 0.302734375,
    "learning_rate": 3.2250000000000005e-06,
    "num_tokens": 207550.0,
    "completions/mean_length": 156.5,
    "completions/min_length": 117.0,
    "completions/max_length": 177.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 156.5,
    "completions/min_terminated_length": 117.0,
    "completions/max_terminated_length": 177.0,
    "rewards/reward_fn/mean": 0.794374942779541,
    "rewards/reward_fn/std": 0.31834450364112854,
    "reward": 0.794374942779541,
    "reward_std": 0.3183445334434509,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.1953434944152832,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 26.463810916997318,
    "epoch": 0.12,
    "step": 72
  },
  {
    "loss": -0.20025718212127686,
    "grad_norm": 0.3359375,
    "learning_rate": 3.2000000000000003e-06,
    "num_tokens": 210661.0,
    "completions/mean_length": 143.75,
    "completions/min_length": 108.0,
    "completions/max_length": 189.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 143.75,
    "completions/min_terminated_length": 108.0,
    "completions/max_terminated_length": 189.0,
    "rewards/reward_fn/mean": 0.6324999928474426,
    "rewards/reward_fn/std": 0.34352341294288635,
    "reward": 0.6324999928474426,
    "reward_std": 0.34352341294288635,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.2760404348373413,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 28.05823507000241,
    "epoch": 0.12166666666666667,
    "step": 73
  },
  {
    "loss": -0.06791014224290848,
    "grad_norm": 0.27734375,
    "learning_rate": 3.175e-06,
    "num_tokens": 213910.0,
    "completions/mean_length": 169.25,
    "completions/min_length": 135.0,
    "completions/max_length": 217.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 169.25,
    "completions/min_terminated_length": 135.0,
    "completions/max_terminated_length": 217.0,
    "rewards/reward_fn/mean": 0.3306249976158142,
    "rewards/reward_fn/std": 0.008750006556510925,
    "reward": 0.3306249976158142,
    "reward_std": 0.008750001899898052,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.1246460676193237,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 31.710728935002408,
    "epoch": 0.12333333333333334,
    "step": 74
  },
  {
    "loss": 0.0,
    "grad_norm": 0.0,
    "learning_rate": 3.1500000000000003e-06,
    "num_tokens": 216608.0,
    "completions/mean_length": 122.5,
    "completions/min_length": 91.0,
    "completions/max_length": 145.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 122.5,
    "completions/min_terminated_length": 91.0,
    "completions/max_terminated_length": 145.0,
    "rewards/reward_fn/mean": 0.33500000834465027,
    "rewards/reward_fn/std": 0.0,
    "reward": 0.33500000834465027,
    "reward_std": 0.0,
    "frac_reward_zero_std": 1.0,
    "entropy": 1.098929762840271,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 21.737217218000296,
    "epoch": 0.125,
    "step": 75
  },
  {
    "loss": 0.14531265199184418,
    "grad_norm": 0.279296875,
    "learning_rate": 3.125e-06,
    "num_tokens": 219434.0,
    "completions/mean_length": 146.5,
    "completions/min_length": 104.0,
    "completions/max_length": 170.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 146.5,
    "completions/min_terminated_length": 104.0,
    "completions/max_terminated_length": 170.0,
    "rewards/reward_fn/mean": 0.5012500286102295,
    "rewards/reward_fn/std": 0.3096066117286682,
    "reward": 0.5012500286102295,
    "reward_std": 0.3096066117286682,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0769822597503662,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 25.090116903000308,
    "epoch": 0.12666666666666668,
    "step": 76
  },
  {
    "loss": -0.00801115296781063,
    "grad_norm": 0.38671875,
    "learning_rate": 3.1000000000000004e-06,
    "num_tokens": 222356.0,
    "completions/mean_length": 121.5,
    "completions/min_length": 113.0,
    "completions/max_length": 131.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 121.5,
    "completions/min_terminated_length": 113.0,
    "completions/max_terminated_length": 131.0,
    "rewards/reward_fn/mean": 0.6657500267028809,
    "rewards/reward_fn/std": 0.3621724843978882,
    "reward": 0.6657500267028809,
    "reward_std": 0.3621724545955658,
    "frac_reward_zero_std": 0.0,
    "entropy": 0.9762334823608398,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 20.045034541999485,
    "epoch": 0.12833333333333333,
    "step": 77
  },
  {
    "loss": -0.055578697472810745,
    "grad_norm": 0.296875,
    "learning_rate": 3.075e-06,
    "num_tokens": 225550.0,
    "completions/mean_length": 124.5,
    "completions/min_length": 77.0,
    "completions/max_length": 165.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 124.5,
    "completions/min_terminated_length": 77.0,
    "completions/max_terminated_length": 165.0,
    "rewards/reward_fn/mean": 0.7949999570846558,
    "rewards/reward_fn/std": 0.30700162053108215,
    "reward": 0.7949999570846558,
    "reward_std": 0.30700162053108215,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.2096550464630127,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 25.065289067002595,
    "epoch": 0.13,
    "step": 78
  },
  {
    "loss": -0.197392076253891,
    "grad_norm": 0.365234375,
    "learning_rate": 3.05e-06,
    "num_tokens": 228395.0,
    "completions/mean_length": 132.25,
    "completions/min_length": 101.0,
    "completions/max_length": 185.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 132.25,
    "completions/min_terminated_length": 101.0,
    "completions/max_terminated_length": 185.0,
    "rewards/reward_fn/mean": 0.48374998569488525,
    "rewards/reward_fn/std": 0.3212572932243347,
    "reward": 0.48374998569488525,
    "reward_std": 0.3212572932243347,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.2732487916946411,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 27.037959209999826,
    "epoch": 0.13166666666666665,
    "step": 79
  },
  {
    "loss": 0.15605716407299042,
    "grad_norm": 0.34765625,
    "learning_rate": 3.0250000000000003e-06,
    "num_tokens": 230893.0,
    "completions/mean_length": 110.5,
    "completions/min_length": 76.0,
    "completions/max_length": 136.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 110.5,
    "completions/min_terminated_length": 76.0,
    "completions/max_terminated_length": 136.0,
    "rewards/reward_fn/mean": 0.4865000247955322,
    "rewards/reward_fn/std": 0.30300000309944153,
    "reward": 0.4865000247955322,
    "reward_std": 0.30300000309944153,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0314747095108032,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 20.13105025200275,
    "epoch": 0.13333333333333333,
    "step": 80
  },
  {
    "loss": 0.0,
    "grad_norm": 0.0,
    "learning_rate": 3e-06,
    "num_tokens": 233867.0,
    "completions/mean_length": 141.5,
    "completions/min_length": 108.0,
    "completions/max_length": 191.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 141.5,
    "completions/min_terminated_length": 108.0,
    "completions/max_terminated_length": 191.0,
    "rewards/reward_fn/mean": 0.33500000834465027,
    "rewards/reward_fn/std": 0.0,
    "reward": 0.33500000834465027,
    "reward_std": 0.0,
    "frac_reward_zero_std": 1.0,
    "entropy": 1.3124823570251465,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 28.160457204001432,
    "epoch": 0.135,
    "step": 81
  },
  {
    "loss": 0.02396450564265251,
    "grad_norm": 0.27734375,
    "learning_rate": 2.9750000000000003e-06,
    "num_tokens": 236651.0,
    "completions/mean_length": 146.0,
    "completions/min_length": 126.0,
    "completions/max_length": 167.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 146.0,
    "completions/min_terminated_length": 126.0,
    "completions/max_terminated_length": 167.0,
    "rewards/reward_fn/mean": 0.8162499666213989,
    "rewards/reward_fn/std": 0.29749998450279236,
    "reward": 0.8162499666213989,
    "reward_std": 0.29749998450279236,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0319082736968994,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 24.461930845001916,
    "epoch": 0.13666666666666666,
    "step": 82
  },
  {
    "loss": 0.06299065053462982,
    "grad_norm": 0.330078125,
    "learning_rate": 2.95e-06,
    "num_tokens": 239411.0,
    "completions/mean_length": 149.0,
    "completions/min_length": 131.0,
    "completions/max_length": 179.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 149.0,
    "completions/min_terminated_length": 131.0,
    "completions/max_terminated_length": 179.0,
    "rewards/reward_fn/mean": 0.5618749856948853,
    "rewards/reward_fn/std": 0.47596120834350586,
    "reward": 0.5618749856948853,
    "reward_std": 0.47596120834350586,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.2378220558166504,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 25.732968376996723,
    "epoch": 0.13833333333333334,
    "step": 83
  },
  {
    "loss": -0.08758159726858139,
    "grad_norm": 0.2470703125,
    "learning_rate": 2.925e-06,
    "num_tokens": 242474.0,
    "completions/mean_length": 172.75,
    "completions/min_length": 142.0,
    "completions/max_length": 191.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 172.75,
    "completions/min_terminated_length": 142.0,
    "completions/max_terminated_length": 191.0,
    "rewards/reward_fn/mean": 0.794374942779541,
    "rewards/reward_fn/std": 0.31834450364112854,
    "reward": 0.794374942779541,
    "reward_std": 0.3183445334434509,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.1778147220611572,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 28.395576832001098,
    "epoch": 0.14,
    "step": 84
  },
  {
    "loss": -0.15402138233184814,
    "grad_norm": 0.294921875,
    "learning_rate": 2.9e-06,
    "num_tokens": 245432.0,
    "completions/mean_length": 127.5,
    "completions/min_length": 88.0,
    "completions/max_length": 156.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 127.5,
    "completions/min_terminated_length": 88.0,
    "completions/max_terminated_length": 156.0,
    "rewards/reward_fn/mean": 0.32625001668930054,
    "rewards/reward_fn/std": 0.017499998211860657,
    "reward": 0.32625001668930054,
    "reward_std": 0.017500003799796104,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.192873477935791,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 23.462746484998206,
    "epoch": 0.14166666666666666,
    "step": 85
  },
  {
    "loss": 0.21816286444664001,
    "grad_norm": 0.326171875,
    "learning_rate": 2.875e-06,
    "num_tokens": 248295.0,
    "completions/mean_length": 135.75,
    "completions/min_length": 103.0,
    "completions/max_length": 195.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 135.75,
    "completions/min_terminated_length": 103.0,
    "completions/max_terminated_length": 195.0,
    "rewards/reward_fn/mean": 0.8075000047683716,
    "rewards/reward_fn/std": 0.3149999976158142,
    "reward": 0.8075000047683716,
    "reward_std": 0.3149999976158142,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.3281114101409912,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 29.33027525199941,
    "epoch": 0.14333333333333334,
    "step": 86
  },
  {
    "loss": -0.050082288682460785,
    "grad_norm": 0.30078125,
    "learning_rate": 2.85e-06,
    "num_tokens": 251030.0,
    "completions/mean_length": 124.75,
    "completions/min_length": 111.0,
    "completions/max_length": 153.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 124.75,
    "completions/min_terminated_length": 111.0,
    "completions/max_terminated_length": 153.0,
    "rewards/reward_fn/mean": 0.9474999904632568,
    "rewards/reward_fn/std": 0.0202072411775589,
    "reward": 0.9474999904632568,
    "reward_std": 0.0202072411775589,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.168349266052246,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 22.54528503599795,
    "epoch": 0.145,
    "step": 87
  },
  {
    "loss": -0.04842241480946541,
    "grad_norm": 0.345703125,
    "learning_rate": 2.825e-06,
    "num_tokens": 253533.0,
    "completions/mean_length": 109.75,
    "completions/min_length": 90.0,
    "completions/max_length": 128.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 109.75,
    "completions/min_terminated_length": 90.0,
    "completions/max_terminated_length": 128.0,
    "rewards/reward_fn/mean": 0.3306249976158142,
    "rewards/reward_fn/std": 0.008750006556510925,
    "reward": 0.3306249976158142,
    "reward_std": 0.008750001899898052,
    "frac_reward_zero_std": 0.0,
    "entropy": 0.9876170754432678,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 18.978010971000913,
    "epoch": 0.14666666666666667,
    "step": 88
  },
  {
    "loss": 0.019071724265813828,
    "grad_norm": 0.314453125,
    "learning_rate": 2.8000000000000003e-06,
    "num_tokens": 256532.0,
    "completions/mean_length": 149.75,
    "completions/min_length": 138.0,
    "completions/max_length": 166.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 149.75,
    "completions/min_terminated_length": 138.0,
    "completions/max_terminated_length": 166.0,
    "rewards/reward_fn/mean": 0.8206250071525574,
    "rewards/reward_fn/std": 0.3005298972129822,
    "reward": 0.8206250071525574,
    "reward_std": 0.3005298972129822,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.289428949356079,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 25.60237361599866,
    "epoch": 0.14833333333333334,
    "step": 89
  },
  {
    "loss": 0.07423283159732819,
    "grad_norm": 0.361328125,
    "learning_rate": 2.7750000000000005e-06,
    "num_tokens": 259055.0,
    "completions/mean_length": 121.75,
    "completions/min_length": 103.0,
    "completions/max_length": 142.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 121.75,
    "completions/min_terminated_length": 103.0,
    "completions/max_terminated_length": 142.0,
    "rewards/reward_fn/mean": 0.2983750104904175,
    "rewards/reward_fn/std": 0.08031539618968964,
    "reward": 0.2983750104904175,
    "reward_std": 0.08031539618968964,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0802006721496582,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 22.080729645000247,
    "epoch": 0.15,
    "step": 90
  },
  {
    "loss": 0.033602241426706314,
    "grad_norm": 0.31640625,
    "learning_rate": 2.7500000000000004e-06,
    "num_tokens": 261798.0,
    "completions/mean_length": 125.75,
    "completions/min_length": 105.0,
    "completions/max_length": 151.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 125.75,
    "completions/min_terminated_length": 105.0,
    "completions/max_terminated_length": 151.0,
    "rewards/reward_fn/mean": 0.8031250238418579,
    "rewards/reward_fn/std": 0.3128456473350525,
    "reward": 0.8031250238418579,
    "reward_std": 0.3128456473350525,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.1015703678131104,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 23.45034698100062,
    "epoch": 0.15166666666666667,
    "step": 91
  },
  {
    "loss": 0.02976590394973755,
    "grad_norm": 0.421875,
    "learning_rate": 2.7250000000000006e-06,
    "num_tokens": 264628.0,
    "completions/mean_length": 108.5,
    "completions/min_length": 83.0,
    "completions/max_length": 145.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 108.5,
    "completions/min_terminated_length": 83.0,
    "completions/max_terminated_length": 145.0,
    "rewards/reward_fn/mean": 0.9728749990463257,
    "rewards/reward_fn/std": 0.015750020742416382,
    "reward": 0.9728749990463257,
    "reward_std": 0.015750011429190636,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.1214567422866821,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 22.66323952300081,
    "epoch": 0.15333333333333332,
    "step": 92
  },
  {
    "loss": 0.0016713386867195368,
    "grad_norm": 0.359375,
    "learning_rate": 2.7000000000000004e-06,
    "num_tokens": 267337.0,
    "completions/mean_length": 119.25,
    "completions/min_length": 101.0,
    "completions/max_length": 146.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 119.25,
    "completions/min_terminated_length": 101.0,
    "completions/max_terminated_length": 146.0,
    "rewards/reward_fn/mean": 0.9772499799728394,
    "rewards/reward_fn/std": 0.015256169252097607,
    "reward": 0.9772499799728394,
    "reward_std": 0.015256185084581375,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.1013864278793335,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 22.21387937399777,
    "epoch": 0.155,
    "step": 93
  },
  {
    "loss": 0.009300082921981812,
    "grad_norm": 0.333984375,
    "learning_rate": 2.6750000000000002e-06,
    "num_tokens": 270352.0,
    "completions/mean_length": 126.75,
    "completions/min_length": 112.0,
    "completions/max_length": 142.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 126.75,
    "completions/min_terminated_length": 112.0,
    "completions/max_terminated_length": 142.0,
    "rewards/reward_fn/mean": 0.48374998569488525,
    "rewards/reward_fn/std": 0.3209393620491028,
    "reward": 0.48374998569488525,
    "reward_std": 0.3209393620491028,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.1367976665496826,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 22.162854308000533,
    "epoch": 0.15666666666666668,
    "step": 94
  },
  {
    "loss": -0.04094283655285835,
    "grad_norm": 0.42578125,
    "learning_rate": 2.6500000000000005e-06,
    "num_tokens": 273143.0,
    "completions/mean_length": 123.75,
    "completions/min_length": 99.0,
    "completions/max_length": 148.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 123.75,
    "completions/min_terminated_length": 99.0,
    "completions/max_terminated_length": 148.0,
    "rewards/reward_fn/mean": 0.9693750143051147,
    "rewards/reward_fn/std": 0.00875002145767212,
    "reward": 0.9693750143051147,
    "reward_std": 0.00875002145767212,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0440210103988647,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 22.38568425199992,
    "epoch": 0.15833333333333333,
    "step": 95
  },
  {
    "loss": 0.15625080466270447,
    "grad_norm": 0.25390625,
    "learning_rate": 2.6250000000000003e-06,
    "num_tokens": 275882.0,
    "completions/mean_length": 140.75,
    "completions/min_length": 97.0,
    "completions/max_length": 160.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 140.75,
    "completions/min_terminated_length": 97.0,
    "completions/max_terminated_length": 160.0,
    "rewards/reward_fn/mean": 0.4925000071525574,
    "rewards/reward_fn/std": 0.2921329438686371,
    "reward": 0.4925000071525574,
    "reward_std": 0.2921329736709595,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.1425111293792725,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 23.621867647998442,
    "epoch": 0.16,
    "step": 96
  },
  {
    "loss": 0.12305271625518799,
    "grad_norm": 0.283203125,
    "learning_rate": 2.6e-06,
    "num_tokens": 278752.0,
    "completions/mean_length": 165.5,
    "completions/min_length": 100.0,
    "completions/max_length": 217.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 165.5,
    "completions/min_terminated_length": 100.0,
    "completions/max_terminated_length": 217.0,
    "rewards/reward_fn/mean": 0.6497499942779541,
    "rewards/reward_fn/std": 0.36401957273483276,
    "reward": 0.6497499942779541,
    "reward_std": 0.3640195429325104,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.1441240310668945,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 32.032986541002174,
    "epoch": 0.16166666666666665,
    "step": 97
  },
  {
    "loss": -0.005260481499135494,
    "grad_norm": 0.302734375,
    "learning_rate": 2.5750000000000003e-06,
    "num_tokens": 280888.0,
    "completions/mean_length": 135.0,
    "completions/min_length": 120.0,
    "completions/max_length": 156.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 135.0,
    "completions/min_terminated_length": 120.0,
    "completions/max_terminated_length": 156.0,
    "rewards/reward_fn/mean": 0.6850000023841858,
    "rewards/reward_fn/std": 0.29560670256614685,
    "reward": 0.6850000023841858,
    "reward_std": 0.29560673236846924,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.078916311264038,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 22.477214457998343,
    "epoch": 0.16333333333333333,
    "step": 98
  },
  {
    "loss": 0.1132018193602562,
    "grad_norm": 0.384765625,
    "learning_rate": 2.55e-06,
    "num_tokens": 283709.0,
    "completions/mean_length": 130.25,
    "completions/min_length": 102.0,
    "completions/max_length": 150.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 130.25,
    "completions/min_terminated_length": 102.0,
    "completions/max_terminated_length": 150.0,
    "rewards/reward_fn/mean": 0.6368750333786011,
    "rewards/reward_fn/std": 0.35040080547332764,
    "reward": 0.6368750333786011,
    "reward_std": 0.35040080547332764,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.333487868309021,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 22.616071093001665,
    "epoch": 0.165,
    "step": 99
  },
  {
    "loss": -0.061634328216314316,
    "grad_norm": 0.421875,
    "learning_rate": 2.5250000000000004e-06,
    "num_tokens": 286963.0,
    "completions/mean_length": 130.5,
    "completions/min_length": 115.0,
    "completions/max_length": 141.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 130.5,
    "completions/min_terminated_length": 115.0,
    "completions/max_terminated_length": 141.0,
    "rewards/reward_fn/mean": 0.3062500059604645,
    "rewards/reward_fn/std": 0.08250000327825546,
    "reward": 0.3062500059604645,
    "reward_std": 0.08250000327825546,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.2107280492782593,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 21.43296745800035,
    "epoch": 0.16666666666666666,
    "step": 100
  },
  {
    "loss": -0.01924452930688858,
    "grad_norm": 0.2734375,
    "learning_rate": 2.5e-06,
    "num_tokens": 289957.0,
    "completions/mean_length": 167.5,
    "completions/min_length": 130.0,
    "completions/max_length": 190.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 167.5,
    "completions/min_terminated_length": 130.0,
    "completions/max_terminated_length": 190.0,
    "rewards/reward_fn/mean": 0.7856249809265137,
    "rewards/reward_fn/std": 0.30188554525375366,
    "reward": 0.7856249809265137,
    "reward_std": 0.30188557505607605,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.2127794027328491,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 28.20487242300078,
    "epoch": 0.16833333333333333,
    "step": 101
  },
  {
    "loss": 0.014912793412804604,
    "grad_norm": 0.365234375,
    "learning_rate": 2.475e-06,
    "num_tokens": 292740.0,
    "completions/mean_length": 163.75,
    "completions/min_length": 106.0,
    "completions/max_length": 247.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 163.75,
    "completions/min_terminated_length": 106.0,
    "completions/max_terminated_length": 247.0,
    "rewards/reward_fn/mean": 0.437250018119812,
    "rewards/reward_fn/std": 0.21189679205417633,
    "reward": 0.437250018119812,
    "reward_std": 0.21189679205417633,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.3349406719207764,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 35.148211256997456,
    "epoch": 0.17,
    "step": 102
  },
  {
    "loss": 0.09064196795225143,
    "grad_norm": 0.310546875,
    "learning_rate": 2.4500000000000003e-06,
    "num_tokens": 295698.0,
    "completions/mean_length": 138.5,
    "completions/min_length": 123.0,
    "completions/max_length": 156.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 138.5,
    "completions/min_terminated_length": 123.0,
    "completions/max_terminated_length": 156.0,
    "rewards/reward_fn/mean": 0.6499999761581421,
    "rewards/reward_fn/std": 0.36373066902160645,
    "reward": 0.6499999761581421,
    "reward_std": 0.36373066902160645,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.2255961894989014,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 23.0402411569994,
    "epoch": 0.17166666666666666,
    "step": 103
  },
  {
    "loss": -0.10342099517583847,
    "grad_norm": 0.298828125,
    "learning_rate": 2.425e-06,
    "num_tokens": 298992.0,
    "completions/mean_length": 183.5,
    "completions/min_length": 122.0,
    "completions/max_length": 220.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 183.5,
    "completions/min_terminated_length": 122.0,
    "completions/max_terminated_length": 220.0,
    "rewards/reward_fn/mean": 0.47462502121925354,
    "rewards/reward_fn/std": 0.29103362560272217,
    "reward": 0.47462502121925354,
    "reward_std": 0.2910335958003998,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.368587613105774,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 32.10494950900102,
    "epoch": 0.17333333333333334,
    "step": 104
  },
  {
    "loss": 0.1281687319278717,
    "grad_norm": 0.39453125,
    "learning_rate": 2.4000000000000003e-06,
    "num_tokens": 301769.0,
    "completions/mean_length": 133.25,
    "completions/min_length": 92.0,
    "completions/max_length": 172.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 133.25,
    "completions/min_terminated_length": 92.0,
    "completions/max_terminated_length": 172.0,
    "rewards/reward_fn/mean": 0.9650000333786011,
    "rewards/reward_fn/std": 0.02474874258041382,
    "reward": 0.9650000333786011,
    "reward_std": 0.02474874258041382,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0861597061157227,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 25.204073330998654,
    "epoch": 0.175,
    "step": 105
  },
  {
    "loss": 0.0,
    "grad_norm": 0.0,
    "learning_rate": 2.375e-06,
    "num_tokens": 304607.0,
    "completions/mean_length": 146.5,
    "completions/min_length": 111.0,
    "completions/max_length": 173.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 146.5,
    "completions/min_terminated_length": 111.0,
    "completions/max_terminated_length": 173.0,
    "rewards/reward_fn/mean": 0.33500000834465027,
    "rewards/reward_fn/std": 0.0,
    "reward": 0.33500000834465027,
    "reward_std": 0.0,
    "frac_reward_zero_std": 1.0,
    "entropy": 1.1364024877548218,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 25.721460639000725,
    "epoch": 0.17666666666666667,
    "step": 106
  },
  {
    "loss": -0.1601293683052063,
    "grad_norm": 0.248046875,
    "learning_rate": 2.35e-06,
    "num_tokens": 307800.0,
    "completions/mean_length": 159.25,
    "completions/min_length": 132.0,
    "completions/max_length": 210.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 159.25,
    "completions/min_terminated_length": 132.0,
    "completions/max_terminated_length": 210.0,
    "rewards/reward_fn/mean": 0.4881249964237213,
    "rewards/reward_fn/std": 0.318023681640625,
    "reward": 0.4881249964237213,
    "reward_std": 0.3180236518383026,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.2100403308868408,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 31.79101676999926,
    "epoch": 0.17833333333333334,
    "step": 107
  },
  {
    "loss": 0.05961640179157257,
    "grad_norm": 0.345703125,
    "learning_rate": 2.325e-06,
    "num_tokens": 310517.0,
    "completions/mean_length": 140.25,
    "completions/min_length": 117.0,
    "completions/max_length": 183.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 140.25,
    "completions/min_terminated_length": 117.0,
    "completions/max_terminated_length": 183.0,
    "rewards/reward_fn/mean": 0.32625001668930054,
    "rewards/reward_fn/std": 0.010103637352585793,
    "reward": 0.32625001668930054,
    "reward_std": 0.010103637352585793,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.120354175567627,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 27.117513699002302,
    "epoch": 0.18,
    "step": 108
  },
  {
    "loss": 0.008527176454663277,
    "grad_norm": 0.3046875,
    "learning_rate": 2.3000000000000004e-06,
    "num_tokens": 312970.0,
    "completions/mean_length": 147.25,
    "completions/min_length": 143.0,
    "completions/max_length": 154.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 147.25,
    "completions/min_terminated_length": 143.0,
    "completions/max_terminated_length": 154.0,
    "rewards/reward_fn/mean": 0.6325000524520874,
    "rewards/reward_fn/std": 0.323947012424469,
    "reward": 0.6325000524520874,
    "reward_std": 0.323947012424469,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.1995975971221924,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 22.515986416998203,
    "epoch": 0.18166666666666667,
    "step": 109
  },
  {
    "loss": 0.23347245156764984,
    "grad_norm": 0.369140625,
    "learning_rate": 2.2750000000000002e-06,
    "num_tokens": 315998.0,
    "completions/mean_length": 139.0,
    "completions/min_length": 90.0,
    "completions/max_length": 177.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 139.0,
    "completions/min_terminated_length": 90.0,
    "completions/max_terminated_length": 177.0,
    "rewards/reward_fn/mean": 0.6587499976158142,
    "rewards/reward_fn/std": 0.35391557216644287,
    "reward": 0.6587499976158142,
    "reward_std": 0.35391557216644287,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.2389236688613892,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 26.32129164900107,
    "epoch": 0.18333333333333332,
    "step": 110
  },
  {
    "loss": -0.00834303256124258,
    "grad_norm": 0.294921875,
    "learning_rate": 2.25e-06,
    "num_tokens": 319005.0,
    "completions/mean_length": 131.75,
    "completions/min_length": 114.0,
    "completions/max_length": 148.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 131.75,
    "completions/min_terminated_length": 114.0,
    "completions/max_terminated_length": 148.0,
    "rewards/reward_fn/mean": 0.9798749685287476,
    "rewards/reward_fn/std": 0.010451287031173706,
    "reward": 0.9798749685287476,
    "reward_std": 0.010451287031173706,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0297316312789917,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 23.76591824099887,
    "epoch": 0.185,
    "step": 111
  },
  {
    "loss": -0.04647402837872505,
    "grad_norm": 0.34765625,
    "learning_rate": 2.2250000000000003e-06,
    "num_tokens": 322001.0,
    "completions/mean_length": 117.0,
    "completions/min_length": 96.0,
    "completions/max_length": 132.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 117.0,
    "completions/min_terminated_length": 96.0,
    "completions/max_terminated_length": 132.0,
    "rewards/reward_fn/mean": 0.9693750143051147,
    "rewards/reward_fn/std": 0.00875002145767212,
    "reward": 0.9693750143051147,
    "reward_std": 0.00875002145767212,
    "frac_reward_zero_std": 0.0,
    "entropy": 0.9716232419013977,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 21.865096360997995,
    "epoch": 0.18666666666666668,
    "step": 112
  },
  {
    "loss": -0.06792350113391876,
    "grad_norm": 0.28125,
    "learning_rate": 2.2e-06,
    "num_tokens": 324989.0,
    "completions/mean_length": 142.0,
    "completions/min_length": 103.0,
    "completions/max_length": 194.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 142.0,
    "completions/min_terminated_length": 103.0,
    "completions/max_terminated_length": 194.0,
    "rewards/reward_fn/mean": 0.7699999809265137,
    "rewards/reward_fn/std": 0.29849621653556824,
    "reward": 0.7699999809265137,
    "reward_std": 0.29849621653556824,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.1111323833465576,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 29.331891907997488,
    "epoch": 0.18833333333333332,
    "step": 113
  },
  {
    "loss": 0.041515663266181946,
    "grad_norm": 0.302734375,
    "learning_rate": 2.1750000000000004e-06,
    "num_tokens": 327838.0,
    "completions/mean_length": 122.25,
    "completions/min_length": 97.0,
    "completions/max_length": 155.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 122.25,
    "completions/min_terminated_length": 97.0,
    "completions/max_terminated_length": 155.0,
    "rewards/reward_fn/mean": 0.8206250071525574,
    "rewards/reward_fn/std": 0.3005298972129822,
    "reward": 0.8206250071525574,
    "reward_std": 0.3005298972129822,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0778565406799316,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 24.199242210001103,
    "epoch": 0.19,
    "step": 114
  },
  {
    "loss": -0.036224182695150375,
    "grad_norm": 0.37109375,
    "learning_rate": 2.15e-06,
    "num_tokens": 330963.0,
    "completions/mean_length": 135.25,
    "completions/min_length": 119.0,
    "completions/max_length": 147.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 135.25,
    "completions/min_terminated_length": 119.0,
    "completions/max_terminated_length": 147.0,
    "rewards/reward_fn/mean": 0.4881250262260437,
    "rewards/reward_fn/std": 0.318023681640625,
    "reward": 0.4881250262260437,
    "reward_std": 0.318023681640625,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.2841877937316895,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 22.69130362800206,
    "epoch": 0.19166666666666668,
    "step": 115
  },
  {
    "loss": 0.0,
    "grad_norm": 0.0,
    "learning_rate": 2.125e-06,
    "num_tokens": 333908.0,
    "completions/mean_length": 144.25,
    "completions/min_length": 115.0,
    "completions/max_length": 159.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 144.25,
    "completions/min_terminated_length": 115.0,
    "completions/max_terminated_length": 159.0,
    "rewards/reward_fn/mean": 0.9649999737739563,
    "rewards/reward_fn/std": 0.0,
    "reward": 0.9649999737739563,
    "reward_std": 0.0,
    "frac_reward_zero_std": 1.0,
    "entropy": 1.1361361742019653,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 23.554109609001898,
    "epoch": 0.19333333333333333,
    "step": 116
  },
  {
    "loss": -0.01241180207580328,
    "grad_norm": 0.27734375,
    "learning_rate": 2.1000000000000002e-06,
    "num_tokens": 336735.0,
    "completions/mean_length": 172.75,
    "completions/min_length": 131.0,
    "completions/max_length": 217.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 172.75,
    "completions/min_terminated_length": 131.0,
    "completions/max_terminated_length": 217.0,
    "rewards/reward_fn/mean": 0.6412500143051147,
    "rewards/reward_fn/std": 0.35391557216644287,
    "reward": 0.6412500143051147,
    "reward_std": 0.35391560196876526,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.189446210861206,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 30.631830683996668,
    "epoch": 0.195,
    "step": 117
  },
  {
    "loss": 0.0,
    "grad_norm": 0.0,
    "learning_rate": 2.075e-06,
    "num_tokens": 338998.0,
    "completions/mean_length": 137.75,
    "completions/min_length": 112.0,
    "completions/max_length": 177.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 137.75,
    "completions/min_terminated_length": 112.0,
    "completions/max_terminated_length": 177.0,
    "rewards/reward_fn/mean": 0.9649999737739563,
    "rewards/reward_fn/std": 0.0,
    "reward": 0.9649999737739563,
    "reward_std": 0.0,
    "frac_reward_zero_std": 1.0,
    "entropy": 1.0830141305923462,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 24.650718585002323,
    "epoch": 0.19666666666666666,
    "step": 118
  },
  {
    "loss": 0.018085498362779617,
    "grad_norm": 0.298828125,
    "learning_rate": 2.05e-06,
    "num_tokens": 342026.0,
    "completions/mean_length": 146.0,
    "completions/min_length": 130.0,
    "completions/max_length": 169.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 146.0,
    "completions/min_terminated_length": 130.0,
    "completions/max_terminated_length": 169.0,
    "rewards/reward_fn/mean": 0.8118749856948853,
    "rewards/reward_fn/std": 0.318023681640625,
    "reward": 0.8118749856948853,
    "reward_std": 0.318023681640625,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.2152268886566162,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 25.480774142000882,
    "epoch": 0.19833333333333333,
    "step": 119
  },
  {
    "loss": -0.015368443913757801,
    "grad_norm": 0.27734375,
    "learning_rate": 2.025e-06,
    "num_tokens": 344997.0,
    "completions/mean_length": 170.75,
    "completions/min_length": 159.0,
    "completions/max_length": 180.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 170.75,
    "completions/min_terminated_length": 159.0,
    "completions/max_terminated_length": 180.0,
    "rewards/reward_fn/mean": 0.4925000071525574,
    "rewards/reward_fn/std": 0.3149999678134918,
    "reward": 0.4925000071525574,
    "reward_std": 0.3149999976158142,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.2765536308288574,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 26.798823175002326,
    "epoch": 0.2,
    "step": 120
  },
  {
    "loss": -0.10509224236011505,
    "grad_norm": 0.375,
    "learning_rate": 2.0000000000000003e-06,
    "num_tokens": 347101.0,
    "completions/mean_length": 123.0,
    "completions/min_length": 97.0,
    "completions/max_length": 151.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 123.0,
    "completions/min_terminated_length": 97.0,
    "completions/max_terminated_length": 151.0,
    "rewards/reward_fn/mean": 0.9562499523162842,
    "rewards/reward_fn/std": 0.017499983310699463,
    "reward": 0.9562499523162842,
    "reward_std": 0.017499983310699463,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0496407747268677,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 21.680493171999842,
    "epoch": 0.20166666666666666,
    "step": 121
  },
  {
    "loss": 0.060011204332113266,
    "grad_norm": 0.33984375,
    "learning_rate": 1.975e-06,
    "num_tokens": 350008.0,
    "completions/mean_length": 129.75,
    "completions/min_length": 114.0,
    "completions/max_length": 152.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 129.75,
    "completions/min_terminated_length": 114.0,
    "completions/max_terminated_length": 152.0,
    "rewards/reward_fn/mean": 0.9693750143051147,
    "rewards/reward_fn/std": 0.00875002145767212,
    "reward": 0.9693750143051147,
    "reward_std": 0.00875002145767212,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.2729018926620483,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 22.95859299299991,
    "epoch": 0.20333333333333334,
    "step": 122
  },
  {
    "loss": 0.0,
    "grad_norm": 0.0,
    "learning_rate": 1.9500000000000004e-06,
    "num_tokens": 352667.0,
    "completions/mean_length": 110.75,
    "completions/min_length": 99.0,
    "completions/max_length": 120.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 110.75,
    "completions/min_terminated_length": 99.0,
    "completions/max_terminated_length": 120.0,
    "rewards/reward_fn/mean": 0.9649999737739563,
    "rewards/reward_fn/std": 0.0,
    "reward": 0.9649999737739563,
    "reward_std": 0.0,
    "frac_reward_zero_std": 1.0,
    "entropy": 1.0064460039138794,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 18.586940206998406,
    "epoch": 0.205,
    "step": 123
  },
  {
    "loss": -0.09901498258113861,
    "grad_norm": 0.4375,
    "learning_rate": 1.925e-06,
    "num_tokens": 355204.0,
    "completions/mean_length": 117.25,
    "completions/min_length": 94.0,
    "completions/max_length": 144.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 117.25,
    "completions/min_terminated_length": 94.0,
    "completions/max_terminated_length": 144.0,
    "rewards/reward_fn/mean": 0.29750001430511475,
    "rewards/reward_fn/std": 0.07500001043081284,
    "reward": 0.29750001430511475,
    "reward_std": 0.07500000298023224,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0144492387771606,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 21.89907346399923,
    "epoch": 0.20666666666666667,
    "step": 124
  },
  {
    "loss": 0.01803763210773468,
    "grad_norm": 0.341796875,
    "learning_rate": 1.9000000000000002e-06,
    "num_tokens": 358015.0,
    "completions/mean_length": 130.75,
    "completions/min_length": 118.0,
    "completions/max_length": 149.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 130.75,
    "completions/min_terminated_length": 118.0,
    "completions/max_terminated_length": 149.0,
    "rewards/reward_fn/mean": 0.32625001668930054,
    "rewards/reward_fn/std": 0.010103637352585793,
    "reward": 0.32625001668930054,
    "reward_std": 0.010103637352585793,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.1170016527175903,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 22.859302955999738,
    "epoch": 0.20833333333333334,
    "step": 125
  },
  {
    "loss": 0.1318972259759903,
    "grad_norm": 0.33984375,
    "learning_rate": 1.8750000000000003e-06,
    "num_tokens": 360842.0,
    "completions/mean_length": 122.75,
    "completions/min_length": 90.0,
    "completions/max_length": 148.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 122.75,
    "completions/min_terminated_length": 90.0,
    "completions/max_terminated_length": 148.0,
    "rewards/reward_fn/mean": 0.9693750143051147,
    "rewards/reward_fn/std": 0.00875002145767212,
    "reward": 0.9693750143051147,
    "reward_std": 0.00875002145767212,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.15077543258667,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 23.222802561998833,
    "epoch": 0.21,
    "step": 126
  },
  {
    "loss": -0.06622515618801117,
    "grad_norm": 0.2197265625,
    "learning_rate": 1.85e-06,
    "num_tokens": 364028.0,
    "completions/mean_length": 207.5,
    "completions/min_length": 180.0,
    "completions/max_length": 221.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 207.5,
    "completions/min_terminated_length": 180.0,
    "completions/max_terminated_length": 221.0,
    "rewards/reward_fn/mean": 0.8159999847412109,
    "rewards/reward_fn/std": 0.297333687543869,
    "reward": 0.8159999847412109,
    "reward_std": 0.297333687543869,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.2833231687545776,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 33.3597098350001,
    "epoch": 0.21166666666666667,
    "step": 127
  },
  {
    "loss": 0.0,
    "grad_norm": 0.0,
    "learning_rate": 1.825e-06,
    "num_tokens": 366980.0,
    "completions/mean_length": 115.0,
    "completions/min_length": 91.0,
    "completions/max_length": 155.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 115.0,
    "completions/min_terminated_length": 91.0,
    "completions/max_terminated_length": 155.0,
    "rewards/reward_fn/mean": 0.33500000834465027,
    "rewards/reward_fn/std": 0.0,
    "reward": 0.33500000834465027,
    "reward_std": 0.0,
    "frac_reward_zero_std": 1.0,
    "entropy": 0.9939418435096741,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 23.618680087998655,
    "epoch": 0.21333333333333335,
    "step": 128
  },
  {
    "loss": 0.16602951288223267,
    "grad_norm": 0.318359375,
    "learning_rate": 1.8000000000000001e-06,
    "num_tokens": 369676.0,
    "completions/mean_length": 144.0,
    "completions/min_length": 74.0,
    "completions/max_length": 193.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 144.0,
    "completions/min_terminated_length": 74.0,
    "completions/max_terminated_length": 193.0,
    "rewards/reward_fn/mean": 0.8041249513626099,
    "rewards/reward_fn/std": 0.31281474232673645,
    "reward": 0.8041249513626099,
    "reward_std": 0.31281474232673645,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.1560587882995605,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 28.466979378998076,
    "epoch": 0.215,
    "step": 129
  },
  {
    "loss": -0.006709250155836344,
    "grad_norm": 0.314453125,
    "learning_rate": 1.7750000000000002e-06,
    "num_tokens": 372676.0,
    "completions/mean_length": 149.0,
    "completions/min_length": 112.0,
    "completions/max_length": 204.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 149.0,
    "completions/min_terminated_length": 112.0,
    "completions/max_terminated_length": 204.0,
    "rewards/reward_fn/mean": 0.4925000071525574,
    "rewards/reward_fn/std": 0.3149999678134918,
    "reward": 0.4925000071525574,
    "reward_std": 0.3149999976158142,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.2238303422927856,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 30.414441388998966,
    "epoch": 0.21666666666666667,
    "step": 130
  },
  {
    "loss": -0.0529935248196125,
    "grad_norm": 0.359375,
    "learning_rate": 1.75e-06,
    "num_tokens": 375679.0,
    "completions/mean_length": 160.75,
    "completions/min_length": 129.0,
    "completions/max_length": 195.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 160.75,
    "completions/min_terminated_length": 129.0,
    "completions/max_terminated_length": 195.0,
    "rewards/reward_fn/mean": 0.9772499799728394,
    "rewards/reward_fn/std": 0.015256169252097607,
    "reward": 0.9772499799728394,
    "reward_std": 0.015256185084581375,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.114032506942749,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 28.930023082000844,
    "epoch": 0.21833333333333332,
    "step": 131
  },
  {
    "loss": -0.016852514818310738,
    "grad_norm": 0.30078125,
    "learning_rate": 1.725e-06,
    "num_tokens": 378594.0,
    "completions/mean_length": 155.75,
    "completions/min_length": 130.0,
    "completions/max_length": 187.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 155.75,
    "completions/min_terminated_length": 130.0,
    "completions/max_terminated_length": 187.0,
    "rewards/reward_fn/mean": 0.32362502813339233,
    "rewards/reward_fn/std": 0.01381650660187006,
    "reward": 0.32362502813339233,
    "reward_std": 0.01381650846451521,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0404397249221802,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 27.9692573909997,
    "epoch": 0.22,
    "step": 132
  },
  {
    "loss": -0.021115276962518692,
    "grad_norm": 0.3515625,
    "learning_rate": 1.7000000000000002e-06,
    "num_tokens": 381292.0,
    "completions/mean_length": 129.5,
    "completions/min_length": 106.0,
    "completions/max_length": 158.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 129.5,
    "completions/min_terminated_length": 106.0,
    "completions/max_terminated_length": 158.0,
    "rewards/reward_fn/mean": 0.34375,
    "rewards/reward_fn/std": 0.017499998211860657,
    "reward": 0.34375,
    "reward_std": 0.017500003799796104,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.1616339683532715,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 23.989496837002662,
    "epoch": 0.22166666666666668,
    "step": 133
  },
  {
    "loss": 0.024493159726262093,
    "grad_norm": 0.267578125,
    "learning_rate": 1.6750000000000003e-06,
    "num_tokens": 384255.0,
    "completions/mean_length": 137.75,
    "completions/min_length": 131.0,
    "completions/max_length": 145.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 137.75,
    "completions/min_terminated_length": 131.0,
    "completions/max_terminated_length": 145.0,
    "rewards/reward_fn/mean": 0.4925000071525574,
    "rewards/reward_fn/std": 0.3149999678134918,
    "reward": 0.4925000071525574,
    "reward_std": 0.3149999976158142,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.1078652143478394,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 22.140087828000105,
    "epoch": 0.22333333333333333,
    "step": 134
  },
  {
    "loss": -0.05794854834675789,
    "grad_norm": 0.365234375,
    "learning_rate": 1.6500000000000003e-06,
    "num_tokens": 386962.0,
    "completions/mean_length": 130.75,
    "completions/min_length": 103.0,
    "completions/max_length": 152.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 130.75,
    "completions/min_terminated_length": 103.0,
    "completions/max_terminated_length": 152.0,
    "rewards/reward_fn/mean": 0.9728749990463257,
    "rewards/reward_fn/std": 0.015750020742416382,
    "reward": 0.9728749990463257,
    "reward_std": 0.015750011429190636,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.1481397151947021,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 22.68116860399823,
    "epoch": 0.225,
    "step": 135
  },
  {
    "loss": -0.008061882108449936,
    "grad_norm": 0.330078125,
    "learning_rate": 1.6250000000000001e-06,
    "num_tokens": 389566.0,
    "completions/mean_length": 124.0,
    "completions/min_length": 121.0,
    "completions/max_length": 130.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 124.0,
    "completions/min_terminated_length": 121.0,
    "completions/max_terminated_length": 130.0,
    "rewards/reward_fn/mean": 0.8075000047683716,
    "rewards/reward_fn/std": 0.3149999976158142,
    "reward": 0.8075000047683716,
    "reward_std": 0.3149999976158142,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.2222754955291748,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 19.605333135001274,
    "epoch": 0.22666666666666666,
    "step": 136
  },
  {
    "loss": 0.0,
    "grad_norm": 0.0,
    "learning_rate": 1.6000000000000001e-06,
    "num_tokens": 392290.0,
    "completions/mean_length": 137.0,
    "completions/min_length": 128.0,
    "completions/max_length": 152.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 137.0,
    "completions/min_terminated_length": 128.0,
    "completions/max_terminated_length": 152.0,
    "rewards/reward_fn/mean": 0.33500000834465027,
    "rewards/reward_fn/std": 0.0,
    "reward": 0.33500000834465027,
    "reward_std": 0.0,
    "frac_reward_zero_std": 1.0,
    "entropy": 1.1045880317687988,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 22.16470382499756,
    "epoch": 0.22833333333333333,
    "step": 137
  },
  {
    "loss": 0.07870414108037949,
    "grad_norm": 0.3671875,
    "learning_rate": 1.5750000000000002e-06,
    "num_tokens": 394880.0,
    "completions/mean_length": 129.5,
    "completions/min_length": 105.0,
    "completions/max_length": 153.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 129.5,
    "completions/min_terminated_length": 105.0,
    "completions/max_terminated_length": 153.0,
    "rewards/reward_fn/mean": 0.3062500059604645,
    "rewards/reward_fn/std": 0.08250000327825546,
    "reward": 0.3062500059604645,
    "reward_std": 0.08250001072883606,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.1190224885940552,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 22.143214091000118,
    "epoch": 0.23,
    "step": 138
  },
  {
    "loss": -0.30225688219070435,
    "grad_norm": 0.373046875,
    "learning_rate": 1.5500000000000002e-06,
    "num_tokens": 397699.0,
    "completions/mean_length": 139.75,
    "completions/min_length": 79.0,
    "completions/max_length": 223.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 139.75,
    "completions/min_terminated_length": 79.0,
    "completions/max_terminated_length": 223.0,
    "rewards/reward_fn/mean": 0.47712501883506775,
    "rewards/reward_fn/std": 0.30046311020851135,
    "reward": 0.47712501883506775,
    "reward_std": 0.30046308040618896,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0666494369506836,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 31.786553677000484,
    "epoch": 0.23166666666666666,
    "step": 139
  },
  {
    "loss": -0.03009037859737873,
    "grad_norm": 0.333984375,
    "learning_rate": 1.525e-06,
    "num_tokens": 400412.0,
    "completions/mean_length": 141.25,
    "completions/min_length": 134.0,
    "completions/max_length": 150.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 141.25,
    "completions/min_terminated_length": 134.0,
    "completions/max_terminated_length": 150.0,
    "rewards/reward_fn/mean": 0.4802500009536743,
    "rewards/reward_fn/std": 0.32342347502708435,
    "reward": 0.4802500009536743,
    "reward_std": 0.32342347502708435,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.147861123085022,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 22.017525466002553,
    "epoch": 0.23333333333333334,
    "step": 140
  },
  {
    "loss": 0.15167391300201416,
    "grad_norm": 0.388671875,
    "learning_rate": 1.5e-06,
    "num_tokens": 403278.0,
    "completions/mean_length": 129.5,
    "completions/min_length": 81.0,
    "completions/max_length": 172.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 129.5,
    "completions/min_terminated_length": 81.0,
    "completions/max_terminated_length": 172.0,
    "rewards/reward_fn/mean": 0.8233749866485596,
    "rewards/reward_fn/std": 0.2794349491596222,
    "reward": 0.8233749866485596,
    "reward_std": 0.2794349789619446,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.1366225481033325,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 25.47716004799804,
    "epoch": 0.235,
    "step": 141
  },
  {
    "loss": -0.029183873906731606,
    "grad_norm": 0.3359375,
    "learning_rate": 1.475e-06,
    "num_tokens": 405972.0,
    "completions/mean_length": 133.5,
    "completions/min_length": 123.0,
    "completions/max_length": 143.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 133.5,
    "completions/min_terminated_length": 123.0,
    "completions/max_terminated_length": 143.0,
    "rewards/reward_fn/mean": 0.6499999761581421,
    "rewards/reward_fn/std": 0.36373066902160645,
    "reward": 0.6499999761581421,
    "reward_std": 0.36373066902160645,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.071744680404663,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 21.137350061002508,
    "epoch": 0.23666666666666666,
    "step": 142
  },
  {
    "loss": 0.025060316547751427,
    "grad_norm": 0.3671875,
    "learning_rate": 1.45e-06,
    "num_tokens": 408671.0,
    "completions/mean_length": 121.75,
    "completions/min_length": 97.0,
    "completions/max_length": 151.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 121.75,
    "completions/min_terminated_length": 97.0,
    "completions/max_terminated_length": 151.0,
    "rewards/reward_fn/mean": 0.484624981880188,
    "rewards/reward_fn/std": 0.32059407234191895,
    "reward": 0.484624981880188,
    "reward_std": 0.32059407234191895,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0496933460235596,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 22.204842587998428,
    "epoch": 0.23833333333333334,
    "step": 143
  },
  {
    "loss": 0.1501941978931427,
    "grad_norm": 0.357421875,
    "learning_rate": 1.425e-06,
    "num_tokens": 411356.0,
    "completions/mean_length": 124.25,
    "completions/min_length": 99.0,
    "completions/max_length": 162.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 124.25,
    "completions/min_terminated_length": 99.0,
    "completions/max_terminated_length": 162.0,
    "rewards/reward_fn/mean": 0.3306249976158142,
    "rewards/reward_fn/std": 0.008750006556510925,
    "reward": 0.3306249976158142,
    "reward_std": 0.008750001899898052,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0704416036605835,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 23.816178979999677,
    "epoch": 0.24,
    "step": 144
  },
  {
    "loss": 0.03730739653110504,
    "grad_norm": 0.423828125,
    "learning_rate": 1.4000000000000001e-06,
    "num_tokens": 414443.0,
    "completions/mean_length": 125.75,
    "completions/min_length": 120.0,
    "completions/max_length": 136.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 125.75,
    "completions/min_terminated_length": 120.0,
    "completions/max_terminated_length": 136.0,
    "rewards/reward_fn/mean": 0.918749988079071,
    "rewards/reward_fn/std": 0.07110730558633804,
    "reward": 0.918749988079071,
    "reward_std": 0.07110730558633804,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.1473332643508911,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 20.62395732100049,
    "epoch": 0.24166666666666667,
    "step": 145
  },
  {
    "loss": 0.0,
    "grad_norm": 0.0,
    "learning_rate": 1.3750000000000002e-06,
    "num_tokens": 417223.0,
    "completions/mean_length": 133.0,
    "completions/min_length": 99.0,
    "completions/max_length": 161.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 133.0,
    "completions/min_terminated_length": 99.0,
    "completions/max_terminated_length": 161.0,
    "rewards/reward_fn/mean": 0.9649999737739563,
    "rewards/reward_fn/std": 0.0,
    "reward": 0.9649999737739563,
    "reward_std": 0.0,
    "frac_reward_zero_std": 1.0,
    "entropy": 1.1500812768936157,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 23.47268433299905,
    "epoch": 0.24333333333333335,
    "step": 146
  },
  {
    "loss": 0.08234935998916626,
    "grad_norm": 0.267578125,
    "learning_rate": 1.3500000000000002e-06,
    "num_tokens": 420117.0,
    "completions/mean_length": 130.5,
    "completions/min_length": 109.0,
    "completions/max_length": 150.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 130.5,
    "completions/min_terminated_length": 109.0,
    "completions/max_terminated_length": 150.0,
    "rewards/reward_fn/mean": 0.4925000071525574,
    "rewards/reward_fn/std": 0.3149999678134918,
    "reward": 0.4925000071525574,
    "reward_std": 0.3149999976158142,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.165224313735962,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 22.157199847999436,
    "epoch": 0.245,
    "step": 147
  },
  {
    "loss": -0.060243088752031326,
    "grad_norm": 0.32421875,
    "learning_rate": 1.3250000000000002e-06,
    "num_tokens": 423030.0,
    "completions/mean_length": 127.25,
    "completions/min_length": 82.0,
    "completions/max_length": 156.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 127.25,
    "completions/min_terminated_length": 82.0,
    "completions/max_terminated_length": 156.0,
    "rewards/reward_fn/mean": 0.4741249978542328,
    "rewards/reward_fn/std": 0.32808491587638855,
    "reward": 0.4741249978542328,
    "reward_std": 0.32808494567871094,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0290559530258179,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 23.24264987100105,
    "epoch": 0.24666666666666667,
    "step": 148
  },
  {
    "loss": 0.11729130893945694,
    "grad_norm": 0.365234375,
    "learning_rate": 1.3e-06,
    "num_tokens": 425636.0,
    "completions/mean_length": 138.5,
    "completions/min_length": 106.0,
    "completions/max_length": 184.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 138.5,
    "completions/min_terminated_length": 106.0,
    "completions/max_terminated_length": 184.0,
    "rewards/reward_fn/mean": 0.4925000071525574,
    "rewards/reward_fn/std": 0.3149999678134918,
    "reward": 0.4925000071525574,
    "reward_std": 0.3149999976158142,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.1438663005828857,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 27.72107682299975,
    "epoch": 0.24833333333333332,
    "step": 149
  },
  {
    "loss": 0.023503979668021202,
    "grad_norm": 0.33203125,
    "learning_rate": 1.275e-06,
    "num_tokens": 428360.0,
    "completions/mean_length": 160.0,
    "completions/min_length": 128.0,
    "completions/max_length": 203.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 160.0,
    "completions/min_terminated_length": 128.0,
    "completions/max_terminated_length": 203.0,
    "rewards/reward_fn/mean": 0.49687498807907104,
    "rewards/reward_fn/std": 0.3128456473350525,
    "reward": 0.49687498807907104,
    "reward_std": 0.3128456175327301,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.3133922815322876,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 29.582455317002314,
    "epoch": 0.25,
    "step": 150
  },
  {
    "loss": 0.030243342742323875,
    "grad_norm": 0.30078125,
    "learning_rate": 1.25e-06,
    "num_tokens": 431399.0,
    "completions/mean_length": 140.75,
    "completions/min_length": 114.0,
    "completions/max_length": 194.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 140.75,
    "completions/min_terminated_length": 114.0,
    "completions/max_terminated_length": 194.0,
    "rewards/reward_fn/mean": 0.3174999952316284,
    "rewards/reward_fn/std": 0.014288689009845257,
    "reward": 0.3174999952316284,
    "reward_std": 0.014288689009845257,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.092279314994812,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 28.14050171800045,
    "epoch": 0.25166666666666665,
    "step": 151
  },
  {
    "loss": -0.11270540952682495,
    "grad_norm": 0.353515625,
    "learning_rate": 1.2250000000000001e-06,
    "num_tokens": 434015.0,
    "completions/mean_length": 141.0,
    "completions/min_length": 120.0,
    "completions/max_length": 172.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 141.0,
    "completions/min_terminated_length": 120.0,
    "completions/max_terminated_length": 172.0,
    "rewards/reward_fn/mean": 0.4881249964237213,
    "rewards/reward_fn/std": 0.2716260850429535,
    "reward": 0.4881249964237213,
    "reward_std": 0.2716260850429535,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.3847908973693848,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 25.02353806499741,
    "epoch": 0.25333333333333335,
    "step": 152
  },
  {
    "loss": 0.15968742966651917,
    "grad_norm": 0.291015625,
    "learning_rate": 1.2000000000000002e-06,
    "num_tokens": 436210.0,
    "completions/mean_length": 138.75,
    "completions/min_length": 93.0,
    "completions/max_length": 180.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 138.75,
    "completions/min_terminated_length": 93.0,
    "completions/max_terminated_length": 180.0,
    "rewards/reward_fn/mean": 0.48375001549720764,
    "rewards/reward_fn/std": 0.27466267347335815,
    "reward": 0.48375001549720764,
    "reward_std": 0.27466264367103577,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.2065187692642212,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 25.63544118199934,
    "epoch": 0.255,
    "step": 153
  },
  {
    "loss": -0.030262991786003113,
    "grad_norm": 0.32421875,
    "learning_rate": 1.175e-06,
    "num_tokens": 439040.0,
    "completions/mean_length": 122.5,
    "completions/min_length": 109.0,
    "completions/max_length": 135.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 122.5,
    "completions/min_terminated_length": 109.0,
    "completions/max_terminated_length": 135.0,
    "rewards/reward_fn/mean": 0.9693750143051147,
    "rewards/reward_fn/std": 0.00875002145767212,
    "reward": 0.9693750143051147,
    "reward_std": 0.00875002145767212,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0571459531784058,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 20.32930038700215,
    "epoch": 0.25666666666666665,
    "step": 154
  },
  {
    "loss": 0.0,
    "grad_norm": 0.0,
    "learning_rate": 1.1500000000000002e-06,
    "num_tokens": 441942.0,
    "completions/mean_length": 135.5,
    "completions/min_length": 114.0,
    "completions/max_length": 155.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 135.5,
    "completions/min_terminated_length": 114.0,
    "completions/max_terminated_length": 155.0,
    "rewards/reward_fn/mean": 0.33500000834465027,
    "rewards/reward_fn/std": 0.0,
    "reward": 0.33500000834465027,
    "reward_std": 0.0,
    "frac_reward_zero_std": 1.0,
    "entropy": 1.005255937576294,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 23.203776197999105,
    "epoch": 0.25833333333333336,
    "step": 155
  },
  {
    "loss": -0.035824455320835114,
    "grad_norm": 0.314453125,
    "learning_rate": 1.125e-06,
    "num_tokens": 444708.0,
    "completions/mean_length": 145.5,
    "completions/min_length": 133.0,
    "completions/max_length": 156.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 145.5,
    "completions/min_terminated_length": 133.0,
    "completions/max_terminated_length": 156.0,
    "rewards/reward_fn/mean": 0.972000002861023,
    "rewards/reward_fn/std": 0.013999998569488525,
    "reward": 0.972000002861023,
    "reward_std": 0.013999998569488525,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0427701473236084,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 23.09061835200191,
    "epoch": 0.26,
    "step": 156
  },
  {
    "loss": 0.031955230981111526,
    "grad_norm": 0.291015625,
    "learning_rate": 1.1e-06,
    "num_tokens": 447946.0,
    "completions/mean_length": 175.5,
    "completions/min_length": 155.0,
    "completions/max_length": 202.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 175.5,
    "completions/min_terminated_length": 155.0,
    "completions/max_terminated_length": 202.0,
    "rewards/reward_fn/mean": 0.5012500286102295,
    "rewards/reward_fn/std": 0.3096066117286682,
    "reward": 0.5012500286102295,
    "reward_std": 0.3096066117286682,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.2932279109954834,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 30.293572720001976,
    "epoch": 0.26166666666666666,
    "step": 157
  },
  {
    "loss": -0.007350577507168055,
    "grad_norm": 0.32421875,
    "learning_rate": 1.075e-06,
    "num_tokens": 450582.0,
    "completions/mean_length": 136.0,
    "completions/min_length": 109.0,
    "completions/max_length": 177.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 136.0,
    "completions/min_terminated_length": 109.0,
    "completions/max_terminated_length": 177.0,
    "rewards/reward_fn/mean": 0.4925000071525574,
    "rewards/reward_fn/std": 0.3149999678134918,
    "reward": 0.4925000071525574,
    "reward_std": 0.3149999976158142,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.183039665222168,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 26.09575506499823,
    "epoch": 0.2633333333333333,
    "step": 158
  },
  {
    "loss": -0.083620585501194,
    "grad_norm": 0.306640625,
    "learning_rate": 1.0500000000000001e-06,
    "num_tokens": 453231.0,
    "completions/mean_length": 132.25,
    "completions/min_length": 110.0,
    "completions/max_length": 151.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 132.25,
    "completions/min_terminated_length": 110.0,
    "completions/max_terminated_length": 151.0,
    "rewards/reward_fn/mean": 0.8162499666213989,
    "rewards/reward_fn/std": 0.32093939185142517,
    "reward": 0.8162499666213989,
    "reward_std": 0.3209393620491028,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0820621252059937,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 22.651842983003007,
    "epoch": 0.265,
    "step": 159
  },
  {
    "loss": -0.023693839088082314,
    "grad_norm": 0.3125,
    "learning_rate": 1.025e-06,
    "num_tokens": 456406.0,
    "completions/mean_length": 137.75,
    "completions/min_length": 101.0,
    "completions/max_length": 185.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 137.75,
    "completions/min_terminated_length": 101.0,
    "completions/max_terminated_length": 185.0,
    "rewards/reward_fn/mean": 0.8162499666213989,
    "rewards/reward_fn/std": 0.32093939185142517,
    "reward": 0.8162499666213989,
    "reward_std": 0.32093939185142517,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.2464767694473267,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 27.780609208999522,
    "epoch": 0.26666666666666666,
    "step": 160
  },
  {
    "loss": 0.1468632072210312,
    "grad_norm": 0.326171875,
    "learning_rate": 1.0000000000000002e-06,
    "num_tokens": 459585.0,
    "completions/mean_length": 154.75,
    "completions/min_length": 107.0,
    "completions/max_length": 201.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 154.75,
    "completions/min_terminated_length": 107.0,
    "completions/max_terminated_length": 201.0,
    "rewards/reward_fn/mean": 0.6587499976158142,
    "rewards/reward_fn/std": 0.3738343119621277,
    "reward": 0.6587499976158142,
    "reward_std": 0.3738343417644501,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.316744327545166,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 29.671674831999553,
    "epoch": 0.2683333333333333,
    "step": 161
  },
  {
    "loss": -0.060441743582487106,
    "grad_norm": 0.3828125,
    "learning_rate": 9.750000000000002e-07,
    "num_tokens": 462603.0,
    "completions/mean_length": 113.5,
    "completions/min_length": 89.0,
    "completions/max_length": 132.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 113.5,
    "completions/min_terminated_length": 89.0,
    "completions/max_terminated_length": 132.0,
    "rewards/reward_fn/mean": 0.32625001668930054,
    "rewards/reward_fn/std": 0.010103637352585793,
    "reward": 0.32625001668930054,
    "reward_std": 0.010103637352585793,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0979673862457275,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 20.17059165000319,
    "epoch": 0.27,
    "step": 162
  },
  {
    "loss": 0.06180550530552864,
    "grad_norm": 0.27734375,
    "learning_rate": 9.500000000000001e-07,
    "num_tokens": 465256.0,
    "completions/mean_length": 127.25,
    "completions/min_length": 106.0,
    "completions/max_length": 140.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 127.25,
    "completions/min_terminated_length": 106.0,
    "completions/max_terminated_length": 140.0,
    "rewards/reward_fn/mean": 0.6543749570846558,
    "rewards/reward_fn/std": 0.3688516914844513,
    "reward": 0.6543749570846558,
    "reward_std": 0.3688516914844513,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.1192840337753296,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 20.585320939000667,
    "epoch": 0.27166666666666667,
    "step": 163
  },
  {
    "loss": 0.0,
    "grad_norm": 0.0,
    "learning_rate": 9.25e-07,
    "num_tokens": 467961.0,
    "completions/mean_length": 117.25,
    "completions/min_length": 92.0,
    "completions/max_length": 146.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 117.25,
    "completions/min_terminated_length": 92.0,
    "completions/max_terminated_length": 146.0,
    "rewards/reward_fn/mean": 0.33500000834465027,
    "rewards/reward_fn/std": 0.0,
    "reward": 0.33500000834465027,
    "reward_std": 0.0,
    "frac_reward_zero_std": 1.0,
    "entropy": 1.060671091079712,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 21.4030541309985,
    "epoch": 0.2733333333333333,
    "step": 164
  },
  {
    "loss": 0.12060170620679855,
    "grad_norm": 0.23828125,
    "learning_rate": 9.000000000000001e-07,
    "num_tokens": 470142.0,
    "completions/mean_length": 122.25,
    "completions/min_length": 92.0,
    "completions/max_length": 162.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 122.25,
    "completions/min_terminated_length": 92.0,
    "completions/max_terminated_length": 162.0,
    "rewards/reward_fn/mean": 0.5056250095367432,
    "rewards/reward_fn/std": 0.30724838376045227,
    "reward": 0.5056250095367432,
    "reward_std": 0.30724838376045227,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.010462760925293,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 23.291043189001357,
    "epoch": 0.275,
    "step": 165
  },
  {
    "loss": -0.0018487756606191397,
    "grad_norm": 0.291015625,
    "learning_rate": 8.75e-07,
    "num_tokens": 473312.0,
    "completions/mean_length": 171.5,
    "completions/min_length": 109.0,
    "completions/max_length": 237.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 171.5,
    "completions/min_terminated_length": 109.0,
    "completions/max_terminated_length": 237.0,
    "rewards/reward_fn/mean": 0.6617499589920044,
    "rewards/reward_fn/std": 0.35750025510787964,
    "reward": 0.6617499589920044,
    "reward_std": 0.35750025510787964,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.3963382244110107,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 34.14518971199868,
    "epoch": 0.27666666666666667,
    "step": 166
  },
  {
    "loss": 0.13671796023845673,
    "grad_norm": 0.314453125,
    "learning_rate": 8.500000000000001e-07,
    "num_tokens": 476340.0,
    "completions/mean_length": 163.0,
    "completions/min_length": 142.0,
    "completions/max_length": 208.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 163.0,
    "completions/min_terminated_length": 142.0,
    "completions/max_terminated_length": 208.0,
    "rewards/reward_fn/mean": 0.968500018119812,
    "rewards/reward_fn/std": 0.028719918802380562,
    "reward": 0.968500018119812,
    "reward_std": 0.028719913214445114,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.246614933013916,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 30.15575728999829,
    "epoch": 0.2783333333333333,
    "step": 167
  },
  {
    "loss": 0.023372644558548927,
    "grad_norm": 0.291015625,
    "learning_rate": 8.250000000000001e-07,
    "num_tokens": 479566.0,
    "completions/mean_length": 137.5,
    "completions/min_length": 129.0,
    "completions/max_length": 149.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 137.5,
    "completions/min_terminated_length": 129.0,
    "completions/max_terminated_length": 149.0,
    "rewards/reward_fn/mean": 0.9693750143051147,
    "rewards/reward_fn/std": 0.00875002145767212,
    "reward": 0.9693750143051147,
    "reward_std": 0.00875002145767212,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.2037601470947266,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 22.651985214000888,
    "epoch": 0.28,
    "step": 168
  },
  {
    "loss": -0.05584796518087387,
    "grad_norm": 0.337890625,
    "learning_rate": 8.000000000000001e-07,
    "num_tokens": 482692.0,
    "completions/mean_length": 119.5,
    "completions/min_length": 106.0,
    "completions/max_length": 137.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 119.5,
    "completions/min_terminated_length": 106.0,
    "completions/max_terminated_length": 137.0,
    "rewards/reward_fn/mean": 0.3306249976158142,
    "rewards/reward_fn/std": 0.008750006556510925,
    "reward": 0.3306249976158142,
    "reward_std": 0.008750001899898052,
    "frac_reward_zero_std": 0.0,
    "entropy": 0.9957237243652344,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 20.86080199800199,
    "epoch": 0.2816666666666667,
    "step": 169
  },
  {
    "loss": 0.013437695801258087,
    "grad_norm": 0.322265625,
    "learning_rate": 7.750000000000001e-07,
    "num_tokens": 485351.0,
    "completions/mean_length": 112.75,
    "completions/min_length": 89.0,
    "completions/max_length": 140.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 112.75,
    "completions/min_terminated_length": 89.0,
    "completions/max_terminated_length": 140.0,
    "rewards/reward_fn/mean": 0.6324999928474426,
    "rewards/reward_fn/std": 0.34352341294288635,
    "reward": 0.6324999928474426,
    "reward_std": 0.34352341294288635,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0123518705368042,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 20.707748422002624,
    "epoch": 0.2833333333333333,
    "step": 170
  },
  {
    "loss": -0.017979849129915237,
    "grad_norm": 0.330078125,
    "learning_rate": 7.5e-07,
    "num_tokens": 488211.0,
    "completions/mean_length": 141.0,
    "completions/min_length": 113.0,
    "completions/max_length": 174.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 141.0,
    "completions/min_terminated_length": 113.0,
    "completions/max_terminated_length": 174.0,
    "rewards/reward_fn/mean": 0.7612500190734863,
    "rewards/reward_fn/std": 0.2913009822368622,
    "reward": 0.7612500190734863,
    "reward_std": 0.2913009524345398,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.2882589101791382,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 25.314323123999202,
    "epoch": 0.285,
    "step": 171
  },
  {
    "loss": 0.0024343333207070827,
    "grad_norm": 0.330078125,
    "learning_rate": 7.25e-07,
    "num_tokens": 491483.0,
    "completions/mean_length": 137.0,
    "completions/min_length": 80.0,
    "completions/max_length": 178.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 137.0,
    "completions/min_terminated_length": 80.0,
    "completions/max_terminated_length": 178.0,
    "rewards/reward_fn/mean": 0.7942500114440918,
    "rewards/reward_fn/std": 0.30649998784065247,
    "reward": 0.7942500114440918,
    "reward_std": 0.30650001764297485,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.2393285036087036,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 26.337294665998343,
    "epoch": 0.2866666666666667,
    "step": 172
  },
  {
    "loss": -0.1553015410900116,
    "grad_norm": 0.326171875,
    "learning_rate": 7.000000000000001e-07,
    "num_tokens": 494264.0,
    "completions/mean_length": 127.25,
    "completions/min_length": 98.0,
    "completions/max_length": 167.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 127.25,
    "completions/min_terminated_length": 98.0,
    "completions/max_terminated_length": 167.0,
    "rewards/reward_fn/mean": 0.34375,
    "rewards/reward_fn/std": 0.017499998211860657,
    "reward": 0.34375,
    "reward_std": 0.017500003799796104,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0118447542190552,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 24.38981057300043,
    "epoch": 0.28833333333333333,
    "step": 173
  },
  {
    "loss": 0.047720227390527725,
    "grad_norm": 0.328125,
    "learning_rate": 6.750000000000001e-07,
    "num_tokens": 497373.0,
    "completions/mean_length": 132.25,
    "completions/min_length": 103.0,
    "completions/max_length": 147.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 132.25,
    "completions/min_terminated_length": 103.0,
    "completions/max_terminated_length": 147.0,
    "rewards/reward_fn/mean": 0.8118749856948853,
    "rewards/reward_fn/std": 0.2953908145427704,
    "reward": 0.8118749856948853,
    "reward_std": 0.2953908145427704,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.3742601871490479,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 22.009925050002494,
    "epoch": 0.29,
    "step": 174
  },
  {
    "loss": 0.11253699660301208,
    "grad_norm": 0.29296875,
    "learning_rate": 6.5e-07,
    "num_tokens": 500407.0,
    "completions/mean_length": 158.5,
    "completions/min_length": 120.0,
    "completions/max_length": 239.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 158.5,
    "completions/min_terminated_length": 120.0,
    "completions/max_terminated_length": 239.0,
    "rewards/reward_fn/mean": 0.9673749804496765,
    "rewards/reward_fn/std": 0.010765498504042625,
    "reward": 0.9673749804496765,
    "reward_std": 0.010765510611236095,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.1549588441848755,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 34.68952643800003,
    "epoch": 0.2916666666666667,
    "step": 175
  },
  {
    "loss": 0.11750662326812744,
    "grad_norm": 0.33984375,
    "learning_rate": 6.25e-07,
    "num_tokens": 503025.0,
    "completions/mean_length": 142.5,
    "completions/min_length": 99.0,
    "completions/max_length": 176.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 142.5,
    "completions/min_terminated_length": 99.0,
    "completions/max_terminated_length": 176.0,
    "rewards/reward_fn/mean": 0.8075000047683716,
    "rewards/reward_fn/std": 0.3149999976158142,
    "reward": 0.8075000047683716,
    "reward_std": 0.3149999976158142,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.1769404411315918,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 26.404072473997076,
    "epoch": 0.29333333333333333,
    "step": 176
  },
  {
    "loss": -0.14939624071121216,
    "grad_norm": 0.392578125,
    "learning_rate": 6.000000000000001e-07,
    "num_tokens": 505870.0,
    "completions/mean_length": 116.25,
    "completions/min_length": 86.0,
    "completions/max_length": 151.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 116.25,
    "completions/min_terminated_length": 86.0,
    "completions/max_terminated_length": 151.0,
    "rewards/reward_fn/mean": 0.484624981880188,
    "rewards/reward_fn/std": 0.32059407234191895,
    "reward": 0.484624981880188,
    "reward_std": 0.32059407234191895,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.045095682144165,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 23.14946024099845,
    "epoch": 0.295,
    "step": 177
  },
  {
    "loss": 0.1224709302186966,
    "grad_norm": 0.255859375,
    "learning_rate": 5.750000000000001e-07,
    "num_tokens": 508679.0,
    "completions/mean_length": 138.25,
    "completions/min_length": 104.0,
    "completions/max_length": 178.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 138.25,
    "completions/min_terminated_length": 104.0,
    "completions/max_terminated_length": 178.0,
    "rewards/reward_fn/mean": 0.3218750059604645,
    "rewards/reward_fn/std": 0.008750006556510925,
    "reward": 0.3218750059604645,
    "reward_std": 0.008750011213123798,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0988469123840332,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 26.35802866800077,
    "epoch": 0.2966666666666667,
    "step": 178
  },
  {
    "loss": 0.007333819754421711,
    "grad_norm": 0.3125,
    "learning_rate": 5.5e-07,
    "num_tokens": 511573.0,
    "completions/mean_length": 124.5,
    "completions/min_length": 107.0,
    "completions/max_length": 150.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 124.5,
    "completions/min_terminated_length": 107.0,
    "completions/max_terminated_length": 150.0,
    "rewards/reward_fn/mean": 0.48375001549720764,
    "rewards/reward_fn/std": 0.3209393620491028,
    "reward": 0.48375001549720764,
    "reward_std": 0.3209393620491028,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0090521574020386,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 22.760440566002217,
    "epoch": 0.29833333333333334,
    "step": 179
  },
  {
    "loss": 0.03865119442343712,
    "grad_norm": 0.3515625,
    "learning_rate": 5.250000000000001e-07,
    "num_tokens": 514181.0,
    "completions/mean_length": 112.0,
    "completions/min_length": 91.0,
    "completions/max_length": 143.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 112.0,
    "completions/min_terminated_length": 91.0,
    "completions/max_terminated_length": 143.0,
    "rewards/reward_fn/mean": 0.6499999761581421,
    "rewards/reward_fn/std": 0.36373066902160645,
    "reward": 0.6499999761581421,
    "reward_std": 0.36373066902160645,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.2578190565109253,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 21.571868511000503,
    "epoch": 0.3,
    "step": 180
  },
  {
    "loss": 0.05512908101081848,
    "grad_norm": 0.36328125,
    "learning_rate": 5.000000000000001e-07,
    "num_tokens": 517096.0,
    "completions/mean_length": 124.75,
    "completions/min_length": 106.0,
    "completions/max_length": 146.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 124.75,
    "completions/min_terminated_length": 106.0,
    "completions/max_terminated_length": 146.0,
    "rewards/reward_fn/mean": 0.6412500143051147,
    "rewards/reward_fn/std": 0.35391557216644287,
    "reward": 0.6412500143051147,
    "reward_std": 0.35391560196876526,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.3287837505340576,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 21.837761578000936,
    "epoch": 0.3016666666666667,
    "step": 181
  },
  {
    "loss": -0.10974273085594177,
    "grad_norm": 0.359375,
    "learning_rate": 4.7500000000000006e-07,
    "num_tokens": 520018.0,
    "completions/mean_length": 116.5,
    "completions/min_length": 89.0,
    "completions/max_length": 152.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 116.5,
    "completions/min_terminated_length": 89.0,
    "completions/max_terminated_length": 152.0,
    "rewards/reward_fn/mean": 0.8136249780654907,
    "rewards/reward_fn/std": 0.29675617814064026,
    "reward": 0.8136249780654907,
    "reward_std": 0.29675617814064026,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0346094369888306,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 22.461976462000166,
    "epoch": 0.30333333333333334,
    "step": 182
  },
  {
    "loss": 0.042166367173194885,
    "grad_norm": 0.365234375,
    "learning_rate": 4.5000000000000003e-07,
    "num_tokens": 522491.0,
    "completions/mean_length": 127.25,
    "completions/min_length": 109.0,
    "completions/max_length": 138.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 127.25,
    "completions/min_terminated_length": 109.0,
    "completions/max_terminated_length": 138.0,
    "rewards/reward_fn/mean": 0.3062500059604645,
    "rewards/reward_fn/std": 0.05750000849366188,
    "reward": 0.3062500059604645,
    "reward_std": 0.057500001043081284,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0929187536239624,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 20.686902885001473,
    "epoch": 0.305,
    "step": 183
  },
  {
    "loss": 0.02638399973511696,
    "grad_norm": 0.36328125,
    "learning_rate": 4.2500000000000006e-07,
    "num_tokens": 525490.0,
    "completions/mean_length": 107.75,
    "completions/min_length": 95.0,
    "completions/max_length": 122.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 107.75,
    "completions/min_terminated_length": 95.0,
    "completions/max_terminated_length": 122.0,
    "rewards/reward_fn/mean": 0.9693750143051147,
    "rewards/reward_fn/std": 0.00875002145767212,
    "reward": 0.9693750143051147,
    "reward_std": 0.00875002145767212,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0409756898880005,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 18.76558146499883,
    "epoch": 0.30666666666666664,
    "step": 184
  },
  {
    "loss": -0.023447571322321892,
    "grad_norm": 0.3671875,
    "learning_rate": 4.0000000000000003e-07,
    "num_tokens": 528287.0,
    "completions/mean_length": 136.25,
    "completions/min_length": 120.0,
    "completions/max_length": 159.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 136.25,
    "completions/min_terminated_length": 120.0,
    "completions/max_terminated_length": 159.0,
    "rewards/reward_fn/mean": 0.6631250381469727,
    "rewards/reward_fn/std": 0.3590344488620758,
    "reward": 0.6631250381469727,
    "reward_std": 0.3590344488620758,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.1596969366073608,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 23.50510012199811,
    "epoch": 0.30833333333333335,
    "step": 185
  },
  {
    "loss": -0.08003614097833633,
    "grad_norm": 0.37890625,
    "learning_rate": 3.75e-07,
    "num_tokens": 531576.0,
    "completions/mean_length": 130.25,
    "completions/min_length": 107.0,
    "completions/max_length": 168.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 130.25,
    "completions/min_terminated_length": 107.0,
    "completions/max_terminated_length": 168.0,
    "rewards/reward_fn/mean": 0.9562499523162842,
    "rewards/reward_fn/std": 0.041658345609903336,
    "reward": 0.9562499523162842,
    "reward_std": 0.04165836051106453,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.1854372024536133,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 25.317642065001564,
    "epoch": 0.31,
    "step": 186
  },
  {
    "loss": -0.11360026895999908,
    "grad_norm": 0.31640625,
    "learning_rate": 3.5000000000000004e-07,
    "num_tokens": 534248.0,
    "completions/mean_length": 132.0,
    "completions/min_length": 109.0,
    "completions/max_length": 162.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 132.0,
    "completions/min_terminated_length": 109.0,
    "completions/max_terminated_length": 162.0,
    "rewards/reward_fn/mean": 0.4925000071525574,
    "rewards/reward_fn/std": 0.3149999678134918,
    "reward": 0.4925000071525574,
    "reward_std": 0.3149999976158142,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.1301367282867432,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 24.889422044998355,
    "epoch": 0.31166666666666665,
    "step": 187
  },
  {
    "loss": 0.06846462935209274,
    "grad_norm": 0.298828125,
    "learning_rate": 3.25e-07,
    "num_tokens": 536917.0,
    "completions/mean_length": 137.25,
    "completions/min_length": 84.0,
    "completions/max_length": 170.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 137.25,
    "completions/min_terminated_length": 84.0,
    "completions/max_terminated_length": 170.0,
    "rewards/reward_fn/mean": 0.6456249952316284,
    "rewards/reward_fn/std": 0.35874998569488525,
    "reward": 0.6456249952316284,
    "reward_std": 0.35874998569488525,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.198452115058899,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 24.68997424599729,
    "epoch": 0.31333333333333335,
    "step": 188
  },
  {
    "loss": 0.2711350619792938,
    "grad_norm": 0.375,
    "learning_rate": 3.0000000000000004e-07,
    "num_tokens": 539912.0,
    "completions/mean_length": 140.75,
    "completions/min_length": 99.0,
    "completions/max_length": 249.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 140.75,
    "completions/min_terminated_length": 99.0,
    "completions/max_terminated_length": 249.0,
    "rewards/reward_fn/mean": 0.9668749570846558,
    "rewards/reward_fn/std": 0.01143370196223259,
    "reward": 0.9668749570846558,
    "reward_std": 0.01143370196223259,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.196223497390747,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 35.739375315999496,
    "epoch": 0.315,
    "step": 189
  },
  {
    "loss": 0.0,
    "grad_norm": 0.0,
    "learning_rate": 2.75e-07,
    "num_tokens": 543321.0,
    "completions/mean_length": 147.25,
    "completions/min_length": 131.0,
    "completions/max_length": 173.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 147.25,
    "completions/min_terminated_length": 131.0,
    "completions/max_terminated_length": 173.0,
    "rewards/reward_fn/mean": 0.9649999737739563,
    "rewards/reward_fn/std": 0.0,
    "reward": 0.9649999737739563,
    "reward_std": 0.0,
    "frac_reward_zero_std": 1.0,
    "entropy": 1.1819069385528564,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 26.15833759999805,
    "epoch": 0.31666666666666665,
    "step": 190
  },
  {
    "loss": 0.04181407764554024,
    "grad_norm": 0.365234375,
    "learning_rate": 2.5000000000000004e-07,
    "num_tokens": 546288.0,
    "completions/mean_length": 150.75,
    "completions/min_length": 129.0,
    "completions/max_length": 174.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 150.75,
    "completions/min_terminated_length": 129.0,
    "completions/max_terminated_length": 174.0,
    "rewards/reward_fn/mean": 0.9693750143051147,
    "rewards/reward_fn/std": 0.00875002145767212,
    "reward": 0.9693750143051147,
    "reward_std": 0.00875002145767212,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.2341399192810059,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 25.681773741998768,
    "epoch": 0.31833333333333336,
    "step": 191
  },
  {
    "loss": -0.02491738647222519,
    "grad_norm": 0.294921875,
    "learning_rate": 2.2500000000000002e-07,
    "num_tokens": 549370.0,
    "completions/mean_length": 165.5,
    "completions/min_length": 101.0,
    "completions/max_length": 268.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 165.5,
    "completions/min_terminated_length": 101.0,
    "completions/max_terminated_length": 268.0,
    "rewards/reward_fn/mean": 0.8023749589920044,
    "rewards/reward_fn/std": 0.3117331564426422,
    "reward": 0.8023749589920044,
    "reward_std": 0.3117331564426422,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.2476222515106201,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 38.91904668999996,
    "epoch": 0.32,
    "step": 192
  },
  {
    "loss": -0.012936650775372982,
    "grad_norm": 0.29296875,
    "learning_rate": 2.0000000000000002e-07,
    "num_tokens": 552399.0,
    "completions/mean_length": 125.25,
    "completions/min_length": 115.0,
    "completions/max_length": 135.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 125.25,
    "completions/min_terminated_length": 115.0,
    "completions/max_terminated_length": 135.0,
    "rewards/reward_fn/mean": 0.9474999904632568,
    "rewards/reward_fn/std": 0.034999996423721313,
    "reward": 0.9474999904632568,
    "reward_std": 0.03500000759959221,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.207331895828247,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 21.57561057500061,
    "epoch": 0.32166666666666666,
    "step": 193
  },
  {
    "loss": 0.08878301829099655,
    "grad_norm": 0.375,
    "learning_rate": 1.7500000000000002e-07,
    "num_tokens": 555579.0,
    "completions/mean_length": 125.0,
    "completions/min_length": 110.0,
    "completions/max_length": 145.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 125.0,
    "completions/min_terminated_length": 110.0,
    "completions/max_terminated_length": 145.0,
    "rewards/reward_fn/mean": 0.6631250381469727,
    "rewards/reward_fn/std": 0.3590344488620758,
    "reward": 0.6631250381469727,
    "reward_std": 0.3590344190597534,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.171470046043396,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 22.777231099997152,
    "epoch": 0.3233333333333333,
    "step": 194
  },
  {
    "loss": -0.06446050107479095,
    "grad_norm": 0.359375,
    "learning_rate": 1.5000000000000002e-07,
    "num_tokens": 558548.0,
    "completions/mean_length": 125.25,
    "completions/min_length": 109.0,
    "completions/max_length": 146.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 125.25,
    "completions/min_terminated_length": 109.0,
    "completions/max_terminated_length": 146.0,
    "rewards/reward_fn/mean": 0.32712501287460327,
    "rewards/reward_fn/std": 0.015750005841255188,
    "reward": 0.32712501287460327,
    "reward_std": 0.015750011429190636,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.1294211149215698,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 22.614405715998146,
    "epoch": 0.325,
    "step": 195
  },
  {
    "loss": 0.06022682413458824,
    "grad_norm": 0.36328125,
    "learning_rate": 1.2500000000000002e-07,
    "num_tokens": 561567.0,
    "completions/mean_length": 119.75,
    "completions/min_length": 97.0,
    "completions/max_length": 137.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 119.75,
    "completions/min_terminated_length": 97.0,
    "completions/max_terminated_length": 137.0,
    "rewards/reward_fn/mean": 0.9667500257492065,
    "rewards/reward_fn/std": 0.011608189903199673,
    "reward": 0.9667500257492065,
    "reward_std": 0.011608189903199673,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0905414819717407,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 20.814484424998227,
    "epoch": 0.32666666666666666,
    "step": 196
  },
  {
    "loss": 0.025765087455511093,
    "grad_norm": 0.2734375,
    "learning_rate": 1.0000000000000001e-07,
    "num_tokens": 564581.0,
    "completions/mean_length": 145.5,
    "completions/min_length": 123.0,
    "completions/max_length": 177.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 145.5,
    "completions/min_terminated_length": 123.0,
    "completions/max_terminated_length": 177.0,
    "rewards/reward_fn/mean": 0.8075000047683716,
    "rewards/reward_fn/std": 0.3149999976158142,
    "reward": 0.8075000047683716,
    "reward_std": 0.3149999976158142,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0498154163360596,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 26.312665902001754,
    "epoch": 0.3283333333333333,
    "step": 197
  },
  {
    "loss": 0.010705118998885155,
    "grad_norm": 0.326171875,
    "learning_rate": 7.500000000000001e-08,
    "num_tokens": 567648.0,
    "completions/mean_length": 119.75,
    "completions/min_length": 106.0,
    "completions/max_length": 140.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 119.75,
    "completions/min_terminated_length": 106.0,
    "completions/max_terminated_length": 140.0,
    "rewards/reward_fn/mean": 0.9772499799728394,
    "rewards/reward_fn/std": 0.015256168320775032,
    "reward": 0.9772499799728394,
    "reward_std": 0.015256163664162159,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0515799522399902,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 21.37977665500148,
    "epoch": 0.33,
    "step": 198
  },
  {
    "loss": -0.1868162900209427,
    "grad_norm": 0.328125,
    "learning_rate": 5.0000000000000004e-08,
    "num_tokens": 570646.0,
    "completions/mean_length": 122.5,
    "completions/min_length": 97.0,
    "completions/max_length": 170.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 122.5,
    "completions/min_terminated_length": 97.0,
    "completions/max_terminated_length": 170.0,
    "rewards/reward_fn/mean": 0.34812501072883606,
    "rewards/reward_fn/std": 0.03880372643470764,
    "reward": 0.34812501072883606,
    "reward_std": 0.03880372270941734,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0080755949020386,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 25.808188715000142,
    "epoch": 0.33166666666666667,
    "step": 199
  },
  {
    "loss": 0.09302064031362534,
    "grad_norm": 0.27734375,
    "learning_rate": 2.5000000000000002e-08,
    "num_tokens": 573155.0,
    "completions/mean_length": 118.25,
    "completions/min_length": 96.0,
    "completions/max_length": 144.0,
    "completions/clipped_ratio": 0.0,
    "completions/mean_terminated_length": 118.25,
    "completions/min_terminated_length": 96.0,
    "completions/max_terminated_length": 144.0,
    "rewards/reward_fn/mean": 0.9693750143051147,
    "rewards/reward_fn/std": 0.00875002145767212,
    "reward": 0.9693750143051147,
    "reward_std": 0.00875002145767212,
    "frac_reward_zero_std": 0.0,
    "entropy": 1.0755765438079834,
    "clip_ratio/low_mean": 0.0,
    "clip_ratio/low_min": 0.0,
    "clip_ratio/high_mean": 0.0,
    "clip_ratio/high_max": 0.0,
    "clip_ratio/region_mean": 0.0,
    "step_time": 21.597608337000565,
    "epoch": 0.3333333333333333,
    "step": 200
  },
  {
    "train_runtime": 5010.8646,
    "train_samples_per_second": 0.16,
    "train_steps_per_second": 0.04,
    "total_flos": 0.0,
    "train_loss": 0.0033340772934025154,
    "epoch": 0.3333333333333333,
    "step": 200
  }
]