{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 8e-05,
  "eval_steps": 500,
  "global_step": 4,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2495.0,
      "completions/max_terminated_length": 2495.0,
      "completions/mean_length": 2341.09375,
      "completions/mean_terminated_length": 2341.09375,
      "completions/min_length": 2157.0,
      "completions/min_terminated_length": 2157.0,
      "entropy": 0.6295297034084797,
      "epoch": 2e-05,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 1.3097562789916992,
      "kl": 0.0,
      "learning_rate": 0.0,
      "loss": 0.3789,
      "num_tokens": 101631.0,
      "reward": -0.46510329842567444,
      "reward_std": 0.5682538747787476,
      "rewards/rollout_reward_func/mean": -0.46510329842567444,
      "rewards/rollout_reward_func/std": 0.5920469760894775,
      "sampling/importance_sampling_ratio/max": 3.0,
      "sampling/importance_sampling_ratio/mean": 0.7730730175971985,
      "sampling/importance_sampling_ratio/min": 1.8192803597338152e-17,
      "sampling/sampling_logp_difference/max": 38.54550552368164,
      "sampling/sampling_logp_difference/mean": 1.0014493465423584,
      "step": 1,
      "step_time": 75.44355932800039
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2673.0,
      "completions/max_terminated_length": 2673.0,
      "completions/mean_length": 2399.59375,
      "completions/mean_terminated_length": 2399.59375,
      "completions/min_length": 2132.0,
      "completions/min_terminated_length": 2132.0,
      "entropy": 0.6335931308567524,
      "epoch": 4e-05,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 1.3101922273635864,
      "kl": 0.0,
      "learning_rate": 1.8e-06,
      "loss": 0.1611,
      "num_tokens": 205643.0,
      "reward": -0.570729672908783,
      "reward_std": 0.46239644289016724,
      "rewards/rollout_reward_func/mean": -0.570729672908783,
      "rewards/rollout_reward_func/std": 0.5204209089279175,
      "sampling/importance_sampling_ratio/max": 3.0,
      "sampling/importance_sampling_ratio/mean": 0.7671786546707153,
      "sampling/importance_sampling_ratio/min": 3.6747809732578574e-17,
      "sampling/sampling_logp_difference/max": 37.84245300292969,
      "sampling/sampling_logp_difference/mean": 1.0149199962615967,
      "step": 2,
      "step_time": 72.11773369000014
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2708.0,
      "completions/max_terminated_length": 2708.0,
      "completions/mean_length": 2417.625,
      "completions/mean_terminated_length": 2417.625,
      "completions/min_length": 1621.0,
      "completions/min_terminated_length": 1621.0,
      "entropy": 0.6265235655009747,
      "epoch": 6e-05,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 1.6844843626022339,
      "kl": 0.002009590338275302,
      "learning_rate": 3.6e-06,
      "loss": 0.1226,
      "num_tokens": 309459.0,
      "reward": -0.4209579825401306,
      "reward_std": 0.606694221496582,
      "rewards/rollout_reward_func/mean": -0.4209579825401306,
      "rewards/rollout_reward_func/std": 0.79085773229599,
      "sampling/importance_sampling_ratio/max": 3.0,
      "sampling/importance_sampling_ratio/mean": 0.7607570886611938,
      "sampling/importance_sampling_ratio/min": 4.809209883409827e-17,
      "sampling/sampling_logp_difference/max": 37.57341384887695,
      "sampling/sampling_logp_difference/mean": 1.0065209865570068,
      "step": 3,
      "step_time": 77.18907972300019
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2565.0,
      "completions/max_terminated_length": 2565.0,
      "completions/mean_length": 2351.6875,
      "completions/mean_terminated_length": 2351.6875,
      "completions/min_length": 1949.0,
      "completions/min_terminated_length": 1949.0,
      "entropy": 0.5958993267267942,
      "epoch": 8e-05,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 1.316472053527832,
      "kl": 0.0026191703545919154,
      "learning_rate": 5.4e-06,
      "loss": 0.1934,
      "num_tokens": 412008.0,
      "reward": -0.29841217398643494,
      "reward_std": 0.7643605470657349,
      "rewards/rollout_reward_func/mean": -0.29841217398643494,
      "rewards/rollout_reward_func/std": 0.8671424984931946,
      "sampling/importance_sampling_ratio/max": 3.0,
      "sampling/importance_sampling_ratio/mean": 0.7666546106338501,
      "sampling/importance_sampling_ratio/min": 3.6131335183099954e-17,
      "sampling/sampling_logp_difference/max": 37.859371185302734,
      "sampling/sampling_logp_difference/mean": 1.038360357284546,
      "step": 4,
      "step_time": 72.7726856999999
    }
  ],
  "logging_steps": 1.0,
  "max_steps": 117,
  "num_input_tokens_seen": 412008,
  "num_train_epochs": 1,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}