{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.04,
  "eval_steps": 50,
  "global_step": 100,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0004,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 40.58326721191406,
      "kl": 0.0,
      "learning_rate": 0.0,
      "loss": 0.0,
      "num_tokens": 8880.0,
      "reward": 0.8125,
      "reward_std": 0.26933756470680237,
      "rewards/compute_reward/mean": 0.8125,
      "rewards/compute_reward/std": 0.40311288833618164,
      "step": 1
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0008,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 41.42949676513672,
      "kl": 0.0,
      "learning_rate": 4e-08,
      "loss": 0.0,
      "num_tokens": 17760.0,
      "reward": 0.25,
      "reward_std": 0.25,
      "rewards/compute_reward/mean": 0.25,
      "rewards/compute_reward/std": 0.44721361994743347,
      "step": 2
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0012,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 50.087196350097656,
      "kl": 0.0009548700036248192,
      "learning_rate": 8e-08,
      "loss": 0.0,
      "num_tokens": 26640.0,
      "reward": 0.25,
      "reward_std": 0.39433756470680237,
      "rewards/compute_reward/mean": 0.25,
      "rewards/compute_reward/std": 0.44721361994743347,
      "step": 3
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0016,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 34.29154586791992,
      "kl": 0.0010083172237500548,
      "learning_rate": 1.2e-07,
      "loss": 0.0,
      "num_tokens": 35520.0,
      "reward": 0.25,
      "reward_std": 0.25,
      "rewards/compute_reward/mean": 0.25,
      "rewards/compute_reward/std": 0.44721361994743347,
      "step": 4
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.002,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 52.334266662597656,
      "kl": 0.0011691565159708261,
      "learning_rate": 1.6e-07,
      "loss": 0.0,
      "num_tokens": 44400.0,
      "reward": 0.75,
      "reward_std": 0.39433756470680237,
      "rewards/compute_reward/mean": 0.75,
      "rewards/compute_reward/std": 0.44721361994743347,
      "step": 5
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0024,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 46.0760612487793,
      "kl": 0.001076721484423615,
      "learning_rate": 2e-07,
      "loss": 0.0,
      "num_tokens": 53280.0,
      "reward": 0.375,
      "reward_std": 0.39433756470680237,
      "rewards/compute_reward/mean": 0.375,
      "rewards/compute_reward/std": 0.5,
      "step": 6
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0028,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 43.4420280456543,
      "kl": 0.0012002449948340654,
      "learning_rate": 2.4e-07,
      "loss": 0.0,
      "num_tokens": 62160.0,
      "reward": 0.5625,
      "reward_std": 0.125,
      "rewards/compute_reward/mean": 0.5625,
      "rewards/compute_reward/std": 0.5123475790023804,
      "step": 7
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0032,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 28.63946533203125,
      "kl": 0.0021694210590794683,
      "learning_rate": 2.8e-07,
      "loss": 0.0,
      "num_tokens": 71040.0,
      "reward": 0.4375,
      "reward_std": 0.125,
      "rewards/compute_reward/mean": 0.4375,
      "rewards/compute_reward/std": 0.5123475790023804,
      "step": 8
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0036,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 26.870580673217773,
      "kl": 0.0032091100583784282,
      "learning_rate": 3.2e-07,
      "loss": 0.0,
      "num_tokens": 79920.0,
      "reward": 0.4375,
      "reward_std": 0.125,
      "rewards/compute_reward/mean": 0.4375,
      "rewards/compute_reward/std": 0.5123475790023804,
      "step": 9
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.004,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 26.891817092895508,
      "kl": 0.0060052815824747086,
      "learning_rate": 3.6e-07,
      "loss": 0.0,
      "num_tokens": 88800.0,
      "reward": 0.375,
      "reward_std": 0.14433756470680237,
      "rewards/compute_reward/mean": 0.375,
      "rewards/compute_reward/std": 0.5,
      "step": 10
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0044,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 25.42474365234375,
      "kl": 0.011448211967945099,
      "learning_rate": 4e-07,
      "loss": 0.0,
      "num_tokens": 97680.0,
      "reward": 0.125,
      "reward_std": 0.14433756470680237,
      "rewards/compute_reward/mean": 0.125,
      "rewards/compute_reward/std": 0.3415650427341461,
      "step": 11
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0048,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 44.18297576904297,
      "kl": 0.01896629249677062,
      "learning_rate": 4.3999999999999997e-07,
      "loss": 0.0001,
      "num_tokens": 106560.0,
      "reward": 0.625,
      "reward_std": 0.25,
      "rewards/compute_reward/mean": 0.625,
      "rewards/compute_reward/std": 0.5,
      "step": 12
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0052,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 44.807559967041016,
      "kl": 0.02193804900161922,
      "learning_rate": 4.8e-07,
      "loss": 0.0001,
      "num_tokens": 115440.0,
      "reward": 0.375,
      "reward_std": 0.39433756470680237,
      "rewards/compute_reward/mean": 0.375,
      "rewards/compute_reward/std": 0.5,
      "step": 13
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0056,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 33.35441970825195,
      "kl": 0.021487060468643904,
      "learning_rate": 5.2e-07,
      "loss": 0.0001,
      "num_tokens": 124320.0,
      "reward": 0.3125,
      "reward_std": 0.125,
      "rewards/compute_reward/mean": 0.3125,
      "rewards/compute_reward/std": 0.4787135720252991,
      "step": 14
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.006,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 39.03325653076172,
      "kl": 0.03645531600341201,
      "learning_rate": 5.6e-07,
      "loss": 0.0001,
      "num_tokens": 133200.0,
      "reward": 0.5,
      "reward_std": 0.39433756470680237,
      "rewards/compute_reward/mean": 0.5,
      "rewards/compute_reward/std": 0.5163977742195129,
      "step": 15
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0064,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 39.2026481628418,
      "kl": 0.041529999114573,
      "learning_rate": 6e-07,
      "loss": 0.0002,
      "num_tokens": 142080.0,
      "reward": 0.375,
      "reward_std": 0.25,
      "rewards/compute_reward/mean": 0.375,
      "rewards/compute_reward/std": 0.5,
      "step": 16
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0068,
      "frac_reward_zero_std": 1.0,
      "grad_norm": 0.1584884077310562,
      "kl": 0.05437520984560251,
      "learning_rate": 6.4e-07,
      "loss": 0.0002,
      "num_tokens": 150960.0,
      "reward": 0.25,
      "reward_std": 0.0,
      "rewards/compute_reward/mean": 0.25,
      "rewards/compute_reward/std": 0.44721361994743347,
      "step": 17
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0072,
      "frac_reward_zero_std": 1.0,
      "grad_norm": 0.19170306622982025,
      "kl": 0.050666666589677334,
      "learning_rate": 6.800000000000001e-07,
      "loss": 0.0002,
      "num_tokens": 159840.0,
      "reward": 0.25,
      "reward_std": 0.0,
      "rewards/compute_reward/mean": 0.25,
      "rewards/compute_reward/std": 0.44721361994743347,
      "step": 18
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0076,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 45.9601936340332,
      "kl": 0.05384485796093941,
      "learning_rate": 7.2e-07,
      "loss": 0.0002,
      "num_tokens": 168720.0,
      "reward": 0.3125,
      "reward_std": 0.41367512941360474,
      "rewards/compute_reward/mean": 0.3125,
      "rewards/compute_reward/std": 0.4787135720252991,
      "step": 19
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.008,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 34.677940368652344,
      "kl": 0.054538323543965816,
      "learning_rate": 7.599999999999999e-07,
      "loss": 0.0002,
      "num_tokens": 177600.0,
      "reward": 0.875,
      "reward_std": 0.25,
      "rewards/compute_reward/mean": 0.875,
      "rewards/compute_reward/std": 0.3415650427341461,
      "step": 20
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0084,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 35.03508758544922,
      "kl": 0.05612980853766203,
      "learning_rate": 8e-07,
      "loss": 0.0002,
      "num_tokens": 186480.0,
      "reward": 0.375,
      "reward_std": 0.14433756470680237,
      "rewards/compute_reward/mean": 0.375,
      "rewards/compute_reward/std": 0.5,
      "step": 21
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0088,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 24.550140380859375,
      "kl": 0.06290358863770962,
      "learning_rate": 8.399999999999999e-07,
      "loss": 0.0003,
      "num_tokens": 195360.0,
      "reward": 0.4375,
      "reward_std": 0.125,
      "rewards/compute_reward/mean": 0.4375,
      "rewards/compute_reward/std": 0.5123475790023804,
      "step": 22
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0092,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 28.05525779724121,
      "kl": 0.06588808912783861,
      "learning_rate": 8.799999999999999e-07,
      "loss": 0.0003,
      "num_tokens": 204240.0,
      "reward": 0.1875,
      "reward_std": 0.125,
      "rewards/compute_reward/mean": 0.1875,
      "rewards/compute_reward/std": 0.40311288833618164,
      "step": 23
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0096,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 40.67657470703125,
      "kl": 0.08455395698547363,
      "learning_rate": 9.2e-07,
      "loss": 0.0003,
      "num_tokens": 213120.0,
      "reward": 0.75,
      "reward_std": 0.25,
      "rewards/compute_reward/mean": 0.75,
      "rewards/compute_reward/std": 0.44721361994743347,
      "step": 24
    },
    {
      "completion_length": 504.5,
      "completions/clipped_ratio": 0.875,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 452.0,
      "completions/mean_length": 504.5,
      "completions/mean_terminated_length": 452.0,
      "completions/min_length": 452.0,
      "completions/min_terminated_length": 452.0,
      "epoch": 0.01,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 30.287317276000977,
      "kl": 0.08439960144460201,
      "learning_rate": 9.6e-07,
      "loss": 0.0003,
      "num_tokens": 221880.0,
      "reward": 0.5,
      "reward_std": 0.25,
      "rewards/compute_reward/mean": 0.5,
      "rewards/compute_reward/std": 0.5163977742195129,
      "step": 25
    },
    {
      "completion_length": 488.0,
      "completions/clipped_ratio": 0.9375,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 128.0,
      "completions/mean_length": 488.0,
      "completions/mean_terminated_length": 128.0,
      "completions/min_length": 128.0,
      "completions/min_terminated_length": 128.0,
      "epoch": 0.0104,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 59.74070739746094,
      "kl": 0.10757334344089031,
      "learning_rate": 1e-06,
      "loss": 0.0004,
      "num_tokens": 230376.0,
      "reward": 0.375,
      "reward_std": 0.5,
      "rewards/compute_reward/mean": 0.375,
      "rewards/compute_reward/std": 0.5,
      "step": 26
    },
    {
      "completion_length": 511.25,
      "completions/clipped_ratio": 0.875,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 508.0,
      "completions/mean_length": 511.25,
      "completions/mean_terminated_length": 506.0,
      "completions/min_length": 504.0,
      "completions/min_terminated_length": 504.0,
      "epoch": 0.0108,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 42.04596710205078,
      "kl": 0.10228226706385612,
      "learning_rate": 1.04e-06,
      "loss": 0.0004,
      "num_tokens": 239244.0,
      "reward": 0.3125,
      "reward_std": 0.41367512941360474,
      "rewards/compute_reward/mean": 0.3125,
      "rewards/compute_reward/std": 0.4787135720252991,
      "step": 27
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0112,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 20.101566314697266,
      "kl": 0.09485878236591816,
      "learning_rate": 1.08e-06,
      "loss": 0.0004,
      "num_tokens": 248124.0,
      "reward": 0.5625,
      "reward_std": 0.125,
      "rewards/compute_reward/mean": 0.5625,
      "rewards/compute_reward/std": 0.5123475790023804,
      "step": 28
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0116,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 31.395845413208008,
      "kl": 0.1172705814242363,
      "learning_rate": 1.12e-06,
      "loss": 0.0005,
      "num_tokens": 257004.0,
      "reward": 0.5,
      "reward_std": 0.25,
      "rewards/compute_reward/mean": 0.5,
      "rewards/compute_reward/std": 0.5163977742195129,
      "step": 29
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.012,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 32.50520706176758,
      "kl": 0.10939355939626694,
      "learning_rate": 1.16e-06,
      "loss": 0.0004,
      "num_tokens": 265884.0,
      "reward": 0.5,
      "reward_std": 0.25,
      "rewards/compute_reward/mean": 0.5,
      "rewards/compute_reward/std": 0.5163977742195129,
      "step": 30
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0124,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 23.288822174072266,
      "kl": 0.11976311914622784,
      "learning_rate": 1.2e-06,
      "loss": 0.0005,
      "num_tokens": 274764.0,
      "reward": 0.125,
      "reward_std": 0.14433756470680237,
      "rewards/compute_reward/mean": 0.125,
      "rewards/compute_reward/std": 0.3415650427341461,
      "step": 31
    },
    {
      "completion_length": 507.5625,
      "completions/clipped_ratio": 0.9375,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 441.0,
      "completions/mean_length": 507.5625,
      "completions/mean_terminated_length": 441.0,
      "completions/min_length": 441.0,
      "completions/min_terminated_length": 441.0,
      "epoch": 0.0128,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 17.808048248291016,
      "kl": 0.12108245491981506,
      "learning_rate": 1.24e-06,
      "loss": 0.0005,
      "num_tokens": 283573.0,
      "reward": 0.6875,
      "reward_std": 0.125,
      "rewards/compute_reward/mean": 0.6875,
      "rewards/compute_reward/std": 0.4787135720252991,
      "step": 32
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0132,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 38.9638557434082,
      "kl": 0.13311625644564629,
      "learning_rate": 1.28e-06,
      "loss": 0.0005,
      "num_tokens": 292453.0,
      "reward": 0.75,
      "reward_std": 0.25,
      "rewards/compute_reward/mean": 0.75,
      "rewards/compute_reward/std": 0.44721361994743347,
      "step": 33
    },
    {
      "completion_length": 511.9375,
      "completions/clipped_ratio": 0.9375,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 511.0,
      "completions/mean_length": 511.9375,
      "completions/mean_terminated_length": 511.0,
      "completions/min_length": 511.0,
      "completions/min_terminated_length": 511.0,
      "epoch": 0.0136,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 35.36914825439453,
      "kl": 0.15050986409187317,
      "learning_rate": 1.32e-06,
      "loss": 0.0006,
      "num_tokens": 301332.0,
      "reward": 0.5,
      "reward_std": 0.28867512941360474,
      "rewards/compute_reward/mean": 0.5,
      "rewards/compute_reward/std": 0.5163977742195129,
      "step": 34
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.014,
      "frac_reward_zero_std": 1.0,
      "grad_norm": 0.13711945712566376,
      "kl": 0.1605079583823681,
      "learning_rate": 1.3600000000000001e-06,
      "loss": 0.0006,
      "num_tokens": 310212.0,
      "reward": 0.0,
      "reward_std": 0.0,
      "rewards/compute_reward/mean": 0.0,
      "rewards/compute_reward/std": 0.0,
      "step": 35
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0144,
      "frac_reward_zero_std": 1.0,
      "grad_norm": 0.1120786964893341,
      "kl": 0.1808343380689621,
      "learning_rate": 1.4e-06,
      "loss": 0.0007,
      "num_tokens": 319092.0,
      "reward": 0.25,
      "reward_std": 0.0,
      "rewards/compute_reward/mean": 0.25,
      "rewards/compute_reward/std": 0.44721361994743347,
      "step": 36
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0148,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 25.245079040527344,
      "kl": 0.17078161984682083,
      "learning_rate": 1.44e-06,
      "loss": 0.0007,
      "num_tokens": 327972.0,
      "reward": 0.3125,
      "reward_std": 0.125,
      "rewards/compute_reward/mean": 0.3125,
      "rewards/compute_reward/std": 0.4787135720252991,
      "step": 37
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0152,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 37.300872802734375,
      "kl": 0.17071689665317535,
      "learning_rate": 1.48e-06,
      "loss": 0.0007,
      "num_tokens": 336852.0,
      "reward": 0.25,
      "reward_std": 0.25,
      "rewards/compute_reward/mean": 0.25,
      "rewards/compute_reward/std": 0.44721361994743347,
      "step": 38
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0156,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 34.02109909057617,
      "kl": 0.17260625213384628,
      "learning_rate": 1.5199999999999998e-06,
      "loss": 0.0007,
      "num_tokens": 345732.0,
      "reward": 0.3125,
      "reward_std": 0.26933756470680237,
      "rewards/compute_reward/mean": 0.3125,
      "rewards/compute_reward/std": 0.4787135720252991,
      "step": 39
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.016,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 22.01497459411621,
      "kl": 0.15412141010165215,
      "learning_rate": 1.5599999999999999e-06,
      "loss": 0.0006,
      "num_tokens": 354612.0,
      "reward": 0.3125,
      "reward_std": 0.125,
      "rewards/compute_reward/mean": 0.3125,
      "rewards/compute_reward/std": 0.4787135720252991,
      "step": 40
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0164,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 34.2176399230957,
      "kl": 0.18766384199261665,
      "learning_rate": 1.6e-06,
      "loss": 0.0008,
      "num_tokens": 363492.0,
      "reward": 0.5,
      "reward_std": 0.28867512941360474,
      "rewards/compute_reward/mean": 0.5,
      "rewards/compute_reward/std": 0.5163977742195129,
      "step": 41
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0168,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 3294.3701171875,
      "kl": 7.936608750373125,
      "learning_rate": 1.6399999999999998e-06,
      "loss": 0.0317,
      "num_tokens": 372372.0,
      "reward": 0.3125,
      "reward_std": 0.125,
      "rewards/compute_reward/mean": 0.3125,
      "rewards/compute_reward/std": 0.4787135720252991,
      "step": 42
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0172,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 33.57102966308594,
      "kl": 0.19400563091039658,
      "learning_rate": 1.6799999999999998e-06,
      "loss": 0.0008,
      "num_tokens": 381252.0,
      "reward": 0.125,
      "reward_std": 0.25,
      "rewards/compute_reward/mean": 0.125,
      "rewards/compute_reward/std": 0.3415650427341461,
      "step": 43
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0176,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 29.105810165405273,
      "kl": 0.17841431871056557,
      "learning_rate": 1.7199999999999998e-06,
      "loss": 0.0007,
      "num_tokens": 390132.0,
      "reward": 0.5625,
      "reward_std": 0.26933756470680237,
      "rewards/compute_reward/mean": 0.5625,
      "rewards/compute_reward/std": 0.5123475790023804,
      "step": 44
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.018,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 31.889549255371094,
      "kl": 0.16708053275942802,
      "learning_rate": 1.7599999999999999e-06,
      "loss": 0.0007,
      "num_tokens": 399012.0,
      "reward": 0.25,
      "reward_std": 0.28867512941360474,
      "rewards/compute_reward/mean": 0.25,
      "rewards/compute_reward/std": 0.44721361994743347,
      "step": 45
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0184,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 27.441055297851562,
      "kl": 0.1539367213845253,
      "learning_rate": 1.8e-06,
      "loss": 0.0006,
      "num_tokens": 407892.0,
      "reward": 0.5,
      "reward_std": 0.25,
      "rewards/compute_reward/mean": 0.5,
      "rewards/compute_reward/std": 0.5163977742195129,
      "step": 46
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0188,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 25.701475143432617,
      "kl": 0.17651546001434326,
      "learning_rate": 1.84e-06,
      "loss": 0.0007,
      "num_tokens": 416772.0,
      "reward": 0.3125,
      "reward_std": 0.26933756470680237,
      "rewards/compute_reward/mean": 0.3125,
      "rewards/compute_reward/std": 0.4787135720252991,
      "step": 47
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0192,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 17.01966667175293,
      "kl": 0.16722217947244644,
      "learning_rate": 1.8799999999999998e-06,
      "loss": 0.0007,
      "num_tokens": 425652.0,
      "reward": 0.1875,
      "reward_std": 0.125,
      "rewards/compute_reward/mean": 0.1875,
      "rewards/compute_reward/std": 0.40311288833618164,
      "step": 48
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0196,
      "frac_reward_zero_std": 1.0,
      "grad_norm": 0.09812258183956146,
      "kl": 0.20355528965592384,
      "learning_rate": 1.92e-06,
      "loss": 0.0008,
      "num_tokens": 434532.0,
      "reward": 1.0,
      "reward_std": 0.0,
      "rewards/compute_reward/mean": 1.0,
      "rewards/compute_reward/std": 0.0,
      "step": 49
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.02,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 24.149648666381836,
      "kl": 0.2126757726073265,
      "learning_rate": 1.96e-06,
      "loss": 0.0009,
      "num_tokens": 443412.0,
      "reward": 0.3125,
      "reward_std": 0.26933756470680237,
      "rewards/compute_reward/mean": 0.3125,
      "rewards/compute_reward/std": 0.4787135720252991,
      "step": 50
    },
    {
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.02,
      "eval_loss": 0.0002288773684995249,
      "eval_runtime": 12487.3227,
      "eval_samples_per_second": 0.016,
      "eval_steps_per_second": 0.004,
      "frac_reward_zero_std": 0.58,
      "num_tokens": 554412.0,
      "reward": 0.415,
      "reward_std": 0.21928203105926514,
      "rewards/compute_reward/mean": 0.415,
      "rewards/compute_reward/std": 0.21928203105926514,
      "step": 50
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0204,
      "frac_reward_zero_std": 1.0,
      "grad_norm": 0.10443887859582901,
      "kl": 0.2396080531179905,
      "learning_rate": 2e-06,
      "loss": 0.001,
      "num_tokens": 563292.0,
      "reward": 0.5,
      "reward_std": 0.0,
      "rewards/compute_reward/mean": 0.5,
      "rewards/compute_reward/std": 0.5163977742195129,
      "step": 51
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0208,
      "frac_reward_zero_std": 1.0,
      "grad_norm": 0.07596234977245331,
      "kl": 0.21716779470443726,
      "learning_rate": 1.9999756307053944e-06,
      "loss": 0.0009,
      "num_tokens": 572172.0,
      "reward": 0.75,
      "reward_std": 0.0,
      "rewards/compute_reward/mean": 0.75,
      "rewards/compute_reward/std": 0.44721361994743347,
      "step": 52
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0212,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 7.932524681091309,
      "kl": 0.22818409651517868,
      "learning_rate": 1.999902524009304e-06,
      "loss": 0.2509,
      "num_tokens": 581052.0,
      "reward": 0.3125,
      "reward_std": 0.125,
      "rewards/compute_reward/mean": 0.3125,
      "rewards/compute_reward/std": 0.4787135720252991,
      "step": 53
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0216,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 17.921937942504883,
      "kl": 0.2106623314321041,
      "learning_rate": 1.999780683474845e-06,
      "loss": -0.4991,
      "num_tokens": 589932.0,
      "reward": 0.3125,
      "reward_std": 0.125,
      "rewards/compute_reward/mean": 0.3125,
      "rewards/compute_reward/std": 0.4787135720252991,
      "step": 54
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.022,
      "frac_reward_zero_std": 1.0,
      "grad_norm": 8.580633163452148,
      "kl": 0.24051300063729286,
      "learning_rate": 1.999610115040354e-06,
      "loss": 0.2509,
      "num_tokens": 598812.0,
      "reward": 0.75,
      "reward_std": 0.0,
      "rewards/compute_reward/mean": 0.75,
      "rewards/compute_reward/std": 0.44721361994743347,
      "step": 55
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0224,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 23.697641372680664,
      "kl": 0.2225131280720234,
      "learning_rate": 1.9993908270190957e-06,
      "loss": 0.2174,
      "num_tokens": 607692.0,
      "reward": 0.25,
      "reward_std": 0.28867512941360474,
      "rewards/compute_reward/mean": 0.25,
      "rewards/compute_reward/std": 0.44721361994743347,
      "step": 56
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0228,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 23.44224739074707,
      "kl": 0.21955753862857819,
      "learning_rate": 1.999122830098858e-06,
      "loss": -0.6485,
      "num_tokens": 616572.0,
      "reward": 0.3125,
      "reward_std": 0.26933756470680237,
      "rewards/compute_reward/mean": 0.3125,
      "rewards/compute_reward/std": 0.4787135720252991,
      "step": 57
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0232,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 18.962669372558594,
      "kl": 0.28268108516931534,
      "learning_rate": 1.998806137341434e-06,
      "loss": 0.4341,
      "num_tokens": 625452.0,
      "reward": 0.375,
      "reward_std": 0.14433756470680237,
      "rewards/compute_reward/mean": 0.375,
      "rewards/compute_reward/std": 0.5,
      "step": 58
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0236,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 32.13456726074219,
      "kl": 0.24665188044309616,
      "learning_rate": 1.998440764181981e-06,
      "loss": 0.001,
      "num_tokens": 634332.0,
      "reward": 0.375,
      "reward_std": 0.39433756470680237,
      "rewards/compute_reward/mean": 0.375,
      "rewards/compute_reward/std": 0.5,
      "step": 59
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.024,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 27.285778045654297,
      "kl": 0.25285593420267105,
      "learning_rate": 1.9980267284282714e-06,
      "loss": -0.3739,
      "num_tokens": 643212.0,
      "reward": 0.1875,
      "reward_std": 0.375,
      "rewards/compute_reward/mean": 0.1875,
      "rewards/compute_reward/std": 0.40311288833618164,
      "step": 60
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0244,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 18.179141998291016,
      "kl": 0.26473674178123474,
      "learning_rate": 1.997564050259824e-06,
      "loss": 0.251,
      "num_tokens": 652092.0,
      "reward": 0.5625,
      "reward_std": 0.125,
      "rewards/compute_reward/mean": 0.5625,
      "rewards/compute_reward/std": 0.5123475790023804,
      "step": 61
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0248,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 23.763172149658203,
      "kl": 0.3116937354207039,
      "learning_rate": 1.99705275222692e-06,
      "loss": 0.7511,
      "num_tokens": 660972.0,
      "reward": 0.6875,
      "reward_std": 0.375,
      "rewards/compute_reward/mean": 0.6875,
      "rewards/compute_reward/std": 0.4787135720252991,
      "step": 62
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0252,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 27.499988555908203,
      "kl": 0.32043570652604103,
      "learning_rate": 1.9964928592495045e-06,
      "loss": -0.9985,
      "num_tokens": 669852.0,
      "reward": 0.75,
      "reward_std": 0.25,
      "rewards/compute_reward/mean": 0.75,
      "rewards/compute_reward/std": 0.44721361994743347,
      "step": 63
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0256,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 22.124732971191406,
      "kl": 0.3462621197104454,
      "learning_rate": 1.99588439861597e-06,
      "loss": -0.0566,
      "num_tokens": 678732.0,
      "reward": 0.75,
      "reward_std": 0.28867512941360474,
      "rewards/compute_reward/mean": 0.75,
      "rewards/compute_reward/std": 0.44721361994743347,
      "step": 64
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.026,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 32.59796905517578,
      "kl": 0.38544727861881256,
      "learning_rate": 1.995227399981831e-06,
      "loss": 1.0594,
      "num_tokens": 687612.0,
      "reward": 0.5,
      "reward_std": 0.39433756470680237,
      "rewards/compute_reward/mean": 0.5,
      "rewards/compute_reward/std": 0.5163977742195129,
      "step": 65
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0264,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 41.3852424621582,
      "kl": 0.4055018573999405,
      "learning_rate": 1.994521895368273e-06,
      "loss": -1.1232,
      "num_tokens": 696492.0,
      "reward": 0.5625,
      "reward_std": 0.51933753490448,
      "rewards/compute_reward/mean": 0.5625,
      "rewards/compute_reward/std": 0.5123475790023804,
      "step": 66
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0268,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 28.504070281982422,
      "kl": 0.4306362494826317,
      "learning_rate": 1.9937679191605962e-06,
      "loss": 0.3766,
      "num_tokens": 705372.0,
      "reward": 0.1875,
      "reward_std": 0.26933756470680237,
      "rewards/compute_reward/mean": 0.1875,
      "rewards/compute_reward/std": 0.40311288833618164,
      "step": 67
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0272,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 33.86847686767578,
      "kl": 0.4562436640262604,
      "learning_rate": 1.992965508106537e-06,
      "loss": -0.2146,
      "num_tokens": 714252.0,
      "reward": 0.6875,
      "reward_std": 0.51933753490448,
      "rewards/compute_reward/mean": 0.6875,
      "rewards/compute_reward/std": 0.4787135720252991,
      "step": 68
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0276,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 30.6969051361084,
      "kl": 0.4892871007323265,
      "learning_rate": 1.9921147013144777e-06,
      "loss": 0.7183,
      "num_tokens": 723132.0,
      "reward": 0.5,
      "reward_std": 0.25,
      "rewards/compute_reward/mean": 0.5,
      "rewards/compute_reward/std": 0.5163977742195129,
      "step": 69
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.028,
      "frac_reward_zero_std": 1.0,
      "grad_norm": 16.82533836364746,
      "kl": 0.4661508873105049,
      "learning_rate": 1.9912155402515414e-06,
      "loss": -0.3731,
      "num_tokens": 732012.0,
      "reward": 0.0,
      "reward_std": 0.0,
      "rewards/compute_reward/mean": 0.0,
      "rewards/compute_reward/std": 0.0,
      "step": 70
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0284,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 16.892255783081055,
      "kl": 0.5023886859416962,
      "learning_rate": 1.99026806874157e-06,
      "loss": 0.002,
      "num_tokens": 740892.0,
      "reward": 0.375,
      "reward_std": 0.14433756470680237,
      "rewards/compute_reward/mean": 0.375,
      "rewards/compute_reward/std": 0.5,
      "step": 71
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0288,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 28.076086044311523,
      "kl": 0.4288725033402443,
      "learning_rate": 1.9892723329629885e-06,
      "loss": 0.7181,
      "num_tokens": 749772.0,
      "reward": 0.625,
      "reward_std": 0.39433756470680237,
      "rewards/compute_reward/mean": 0.625,
      "rewards/compute_reward/std": 0.5,
      "step": 72
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0292,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 26.868600845336914,
      "kl": 0.43238623440265656,
      "learning_rate": 1.9882283814465526e-06,
      "loss": -0.9311,
      "num_tokens": 758652.0,
      "reward": 0.4375,
      "reward_std": 0.26933756470680237,
      "rewards/compute_reward/mean": 0.4375,
      "rewards/compute_reward/std": 0.5123475790023804,
      "step": 73
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0296,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 22.524154663085938,
      "kl": 0.4475055858492851,
      "learning_rate": 1.987136265072988e-06,
      "loss": 0.2183,
      "num_tokens": 767532.0,
      "reward": 0.5,
      "reward_std": 0.28867512941360474,
      "rewards/compute_reward/mean": 0.5,
      "rewards/compute_reward/std": 0.5163977742195129,
      "step": 74
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.03,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 26.058460235595703,
      "kl": 0.4978598430752754,
      "learning_rate": 1.985996037070505e-06,
      "loss": -0.123,
      "num_tokens": 776412.0,
      "reward": 0.4375,
      "reward_std": 0.26933756470680237,
      "rewards/compute_reward/mean": 0.4375,
      "rewards/compute_reward/std": 0.5123475790023804,
      "step": 75
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0304,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 25.16863250732422,
      "kl": 0.39492419362068176,
      "learning_rate": 1.984807753012208e-06,
      "loss": 0.2515,
      "num_tokens": 785292.0,
      "reward": 0.625,
      "reward_std": 0.25,
      "rewards/compute_reward/mean": 0.625,
      "rewards/compute_reward/std": 0.5,
      "step": 76
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0308,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 26.64851951599121,
      "kl": 0.37667443603277206,
      "learning_rate": 1.983571470813386e-06,
      "loss": -0.3734,
      "num_tokens": 794172.0,
      "reward": 0.4375,
      "reward_std": 0.125,
      "rewards/compute_reward/mean": 0.4375,
      "rewards/compute_reward/std": 0.5123475790023804,
      "step": 77
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0312,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 23.185409545898438,
      "kl": 0.38686759769916534,
      "learning_rate": 1.9822872507286887e-06,
      "loss": 0.2515,
      "num_tokens": 803052.0,
      "reward": 0.375,
      "reward_std": 0.14433756470680237,
      "rewards/compute_reward/mean": 0.375,
      "rewards/compute_reward/std": 0.5,
      "step": 78
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0316,
      "frac_reward_zero_std": 1.0,
      "grad_norm": 0.08748236298561096,
      "kl": 0.4463553726673126,
      "learning_rate": 1.9809551553491913e-06,
      "loss": 0.0018,
      "num_tokens": 811932.0,
      "reward": 0.25,
      "reward_std": 0.0,
      "rewards/compute_reward/mean": 0.25,
      "rewards/compute_reward/std": 0.44721361994743347,
      "step": 79
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.032,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 24.110332489013672,
      "kl": 0.36003363132476807,
      "learning_rate": 1.979575249599344e-06,
      "loss": -0.215,
      "num_tokens": 820812.0,
      "reward": 0.375,
      "reward_std": 0.39433756470680237,
      "rewards/compute_reward/mean": 0.375,
      "rewards/compute_reward/std": 0.5,
      "step": 80
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0324,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 18.954856872558594,
      "kl": 0.36876556277275085,
      "learning_rate": 1.9781476007338054e-06,
      "loss": 0.093,
      "num_tokens": 829692.0,
      "reward": 0.5625,
      "reward_std": 0.125,
      "rewards/compute_reward/mean": 0.5625,
      "rewards/compute_reward/std": 0.5123475790023804,
      "step": 81
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0328,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 9.058808326721191,
      "kl": 0.403683140873909,
      "learning_rate": 1.9766722783341677e-06,
      "loss": -0.2483,
      "num_tokens": 838572.0,
      "reward": 0.4375,
      "reward_std": 0.125,
      "rewards/compute_reward/mean": 0.4375,
      "rewards/compute_reward/std": 0.5123475790023804,
      "step": 82
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0332,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 23.458297729492188,
      "kl": 0.4015289470553398,
      "learning_rate": 1.975149354305563e-06,
      "loss": 0.6265,
      "num_tokens": 847452.0,
      "reward": 0.4375,
      "reward_std": 0.26933756470680237,
      "rewards/compute_reward/mean": 0.4375,
      "rewards/compute_reward/std": 0.5123475790023804,
      "step": 83
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0336,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 21.95713996887207,
      "kl": 0.3268972262740135,
      "learning_rate": 1.97357890287316e-06,
      "loss": -0.2486,
      "num_tokens": 856332.0,
      "reward": 0.25,
      "reward_std": 0.28867512941360474,
      "rewards/compute_reward/mean": 0.25,
      "rewards/compute_reward/std": 0.44721361994743347,
      "step": 84
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.034,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 28.32386589050293,
      "kl": 0.38899460434913635,
      "learning_rate": 1.9719610005785463e-06,
      "loss": -0.7148,
      "num_tokens": 865212.0,
      "reward": 0.625,
      "reward_std": 0.39433756470680237,
      "rewards/compute_reward/mean": 0.625,
      "rewards/compute_reward/std": 0.5,
      "step": 85
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0344,
      "frac_reward_zero_std": 1.0,
      "grad_norm": 41.013309478759766,
      "kl": 0.5532749891281128,
      "learning_rate": 1.9702957262759963e-06,
      "loss": 0.7186,
      "num_tokens": 874092.0,
      "reward": 0.5,
      "reward_std": 0.0,
      "rewards/compute_reward/mean": 0.5,
      "rewards/compute_reward/std": 0.5163977742195129,
      "step": 86
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0348,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 49.202369689941406,
      "kl": 0.5244565531611443,
      "learning_rate": 1.968583161128631e-06,
      "loss": 0.377,
      "num_tokens": 882972.0,
      "reward": 0.4375,
      "reward_std": 0.375,
      "rewards/compute_reward/mean": 0.4375,
      "rewards/compute_reward/std": 0.5123475790023804,
      "step": 87
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0352,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 456.4792785644531,
      "kl": 3.696738824248314,
      "learning_rate": 1.9668233886044593e-06,
      "loss": -0.0187,
      "num_tokens": 891852.0,
      "reward": 0.5625,
      "reward_std": 0.26933756470680237,
      "rewards/compute_reward/mean": 0.5625,
      "rewards/compute_reward/std": 0.5123475790023804,
      "step": 88
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0356,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 77.5242919921875,
      "kl": 0.61634162068367,
      "learning_rate": 1.9650164944723112e-06,
      "loss": 0.0025,
      "num_tokens": 900732.0,
      "reward": 0.4375,
      "reward_std": 0.41367512941360474,
      "rewards/compute_reward/mean": 0.4375,
      "rewards/compute_reward/std": 0.5123475790023804,
      "step": 89
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.036,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 32.868568420410156,
      "kl": 0.5600514858961105,
      "learning_rate": 1.963162566797658e-06,
      "loss": -0.3392,
      "num_tokens": 909612.0,
      "reward": 0.5,
      "reward_std": 0.28867512941360474,
      "rewards/compute_reward/mean": 0.5,
      "rewards/compute_reward/std": 0.5163977742195129,
      "step": 90
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0364,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 32.02371597290039,
      "kl": 0.5822872817516327,
      "learning_rate": 1.9612616959383188e-06,
      "loss": -0.3391,
      "num_tokens": 918492.0,
      "reward": 0.6875,
      "reward_std": 0.26933756470680237,
      "rewards/compute_reward/mean": 0.6875,
      "rewards/compute_reward/std": 0.4787135720252991,
      "step": 91
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0368,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 44.31180953979492,
      "kl": 0.7215879261493683,
      "learning_rate": 1.9593139745400573e-06,
      "loss": 0.8442,
      "num_tokens": 927372.0,
      "reward": 0.5,
      "reward_std": 0.25,
      "rewards/compute_reward/mean": 0.5,
      "rewards/compute_reward/std": 0.5163977742195129,
      "step": 92
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0372,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 61.617347717285156,
      "kl": 0.5873569175601006,
      "learning_rate": 1.957319497532067e-06,
      "loss": -0.7475,
      "num_tokens": 936252.0,
      "reward": 0.75,
      "reward_std": 0.25,
      "rewards/compute_reward/mean": 0.75,
      "rewards/compute_reward/std": 0.44721361994743347,
      "step": 93
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0376,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 45.146324157714844,
      "kl": 0.5552276074886322,
      "learning_rate": 1.9552783621223435e-06,
      "loss": 0.1272,
      "num_tokens": 945132.0,
      "reward": 0.8125,
      "reward_std": 0.375,
      "rewards/compute_reward/mean": 0.8125,
      "rewards/compute_reward/std": 0.40311288833618164,
      "step": 94
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.038,
      "frac_reward_zero_std": 1.0,
      "grad_norm": 24.78557586669922,
      "kl": 0.5455406308174133,
      "learning_rate": 1.953190667792947e-06,
      "loss": 0.1272,
      "num_tokens": 954012.0,
      "reward": 0.5,
      "reward_std": 0.0,
      "rewards/compute_reward/mean": 0.5,
      "rewards/compute_reward/std": 0.5163977742195129,
      "step": 95
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0384,
      "frac_reward_zero_std": 0.75,
      "grad_norm": 11.901802062988281,
      "kl": 0.6519858688116074,
      "learning_rate": 1.9510565162951534e-06,
      "loss": -0.3723,
      "num_tokens": 962892.0,
      "reward": 0.4375,
      "reward_std": 0.125,
      "rewards/compute_reward/mean": 0.4375,
      "rewards/compute_reward/std": 0.5123475790023804,
      "step": 96
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0388,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 34.490142822265625,
      "kl": 0.6270407885313034,
      "learning_rate": 1.9488760116444964e-06,
      "loss": 0.6274,
      "num_tokens": 971772.0,
      "reward": 0.5625,
      "reward_std": 0.375,
      "rewards/compute_reward/mean": 0.5625,
      "rewards/compute_reward/std": 0.5123475790023804,
      "step": 97
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0392,
      "frac_reward_zero_std": 1.0,
      "grad_norm": 21.449689865112305,
      "kl": 0.6352937817573547,
      "learning_rate": 1.9466492601156963e-06,
      "loss": -0.2474,
      "num_tokens": 980652.0,
      "reward": 0.25,
      "reward_std": 0.0,
      "rewards/compute_reward/mean": 0.25,
      "rewards/compute_reward/std": 0.44721361994743347,
      "step": 98
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0396,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 26.181488037109375,
      "kl": 0.6954683512449265,
      "learning_rate": 1.944376370237481e-06,
      "loss": -0.2472,
      "num_tokens": 989532.0,
      "reward": 0.5,
      "reward_std": 0.25,
      "rewards/compute_reward/mean": 0.5,
      "rewards/compute_reward/std": 0.5163977742195129,
      "step": 99
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.04,
      "frac_reward_zero_std": 0.5,
      "grad_norm": 43.50558853149414,
      "kl": 0.6294843703508377,
      "learning_rate": 1.9420574527872966e-06,
      "loss": 0.6274,
      "num_tokens": 998412.0,
      "reward": 0.25,
      "reward_std": 0.25,
      "rewards/compute_reward/mean": 0.25,
      "rewards/compute_reward/std": 0.44721361994743347,
      "step": 100
    },
    {
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.04,
      "eval_loss": 0.13216377794742584,
      "eval_runtime": 13737.4228,
      "eval_samples_per_second": 0.015,
      "eval_steps_per_second": 0.004,
      "frac_reward_zero_std": 0.62,
      "num_tokens": 1109412.0,
      "reward": 0.43,
      "reward_std": 0.20082903623580933,
      "rewards/compute_reward/mean": 0.43,
      "rewards/compute_reward/std": 0.20082903623580933,
      "step": 100
    }
  ],
  "logging_steps": 1,
  "max_steps": 500,
  "num_input_tokens_seen": 1109412,
  "num_train_epochs": 1,
  "save_steps": 50,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}