{
    "epoch": 3.0,
    "eval_logits/chosen": -2.739661693572998,
    "eval_logits/rejected": -2.6780877113342285,
    "eval_logps/chosen": -339.68280029296875,
    "eval_logps/rejected": -325.6119384765625,
    "eval_loss": 0.6874369978904724,
    "eval_rewards/accuracies": 0.7857142686843872,
    "eval_rewards/chosen": -4.315011501312256,
    "eval_rewards/margins": 3.7554383277893066,
    "eval_rewards/rejected": -8.070449829101562,
    "eval_runtime": 84.3732,
    "eval_samples": 2000,
    "eval_samples_per_second": 23.704,
    "eval_steps_per_second": 0.747
}