{
    "epoch": 2.0,
    "eval_logits/chosen": 1.4759222269058228,
    "eval_logits/rejected": 2.240403652191162,
    "eval_logps/chosen": -421.74468994140625,
    "eval_logps/rejected": -544.37890625,
    "eval_loss": 0.5015926957130432,
    "eval_rewards/accuracies": 0.76953125,
    "eval_rewards/chosen": -1.5911476612091064,
    "eval_rewards/margins": 1.2260199785232544,
    "eval_rewards/rejected": -2.8171677589416504,
    "eval_runtime": 93.5372,
    "eval_samples": 2000,
    "eval_samples_per_second": 21.382,
    "eval_steps_per_second": 0.342
}