{
    "epoch": 1.0,
    "eval_logits/chosen": -1.7276861667633057,
    "eval_logits/rejected": -1.6730353832244873,
    "eval_logps/chosen": -948.7984619140625,
    "eval_logps/rejected": -961.3382568359375,
    "eval_loss": 0.03925349563360214,
    "eval_rewards/accuracies": 1.0,
    "eval_rewards/chosen": -30.879230499267578,
    "eval_rewards/margins": 4.054325103759766,
    "eval_rewards/rejected": -34.933555603027344,
    "eval_runtime": 96.1529,
    "eval_samples": 2000,
    "eval_samples_per_second": 20.8,
    "eval_steps_per_second": 0.333
}