{
    "epoch": 1.0,
    "eval_logits/chosen": 0.7988529205322266,
    "eval_logits/rejected": 1.2523157596588135,
    "eval_logps/chosen": -385.9081726074219,
    "eval_logps/rejected": -457.0127868652344,
    "eval_loss": 0.5133188962936401,
    "eval_rewards/accuracies": 0.75390625,
    "eval_rewards/chosen": -1.244707703590393,
    "eval_rewards/margins": 0.8671280741691589,
    "eval_rewards/rejected": -2.1118357181549072,
    "eval_runtime": 99.9074,
    "eval_samples": 2000,
    "eval_samples_per_second": 20.019,
    "eval_steps_per_second": 0.32,
    "train_loss": 0.5577758540044768,
    "train_runtime": 7516.1301,
    "train_samples": 61155,
    "train_samples_per_second": 8.137,
    "train_steps_per_second": 0.254
}