{
    "epoch": 1.0,
    "eval_logits/chosen": -2.290846109390259,
    "eval_logits/rejected": -2.137458324432373,
    "eval_logps/chosen": -285.9466247558594,
    "eval_logps/rejected": -275.9525146484375,
    "eval_loss": 0.3380393981933594,
    "eval_rewards/accuracies": 0.8789682388305664,
    "eval_rewards/chosen": -0.13388866186141968,
    "eval_rewards/margins": 2.963660955429077,
    "eval_rewards/rejected": -3.0975501537323,
    "eval_runtime": 720.1571,
    "eval_samples": 2000,
    "eval_samples_per_second": 2.777,
    "eval_steps_per_second": 0.087
}