{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.2016,
  "eval_steps": 63,
  "global_step": 63,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.2016,
      "eval_logps/chosen": -106.34315490722656,
      "eval_logps/ref_chosen": -70.265625,
      "eval_logps/ref_rejected": -63.51186752319336,
      "eval_logps/rejected": -106.34315490722656,
      "eval_loss": 0.6557361483573914,
      "eval_rewards/accuracies": 0.6162024772678456,
      "eval_rewards/chosen": -1.8038759231567383,
      "eval_rewards/margins": 0.1433669477701187,
      "eval_rewards/rejected": -1.9472428560256958,
      "eval_runtime": 2599.3441,
      "eval_samples_per_second": 32.24,
      "eval_steps_per_second": 4.03,
      "step": 63
    }
  ],
  "logging_steps": 100,
  "max_steps": 312,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 63,
  "total_flos": 0.0,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}