{
    "epoch": 1.0,
    "eval_logits/chosen": -0.9365912079811096,
    "eval_logits/rejected": -0.926014244556427,
    "eval_logps/chosen": -352.8179626464844,
    "eval_logps/rejected": -349.7088317871094,
    "eval_loss": 0.18906600773334503,
    "eval_pred_label": 5005.27001953125,
    "eval_rewards/accuracies": 0.7579365372657776,
    "eval_rewards/chosen": 0.2332553267478943,
    "eval_rewards/margins": 3.972490072250366,
    "eval_rewards/rejected": -3.7392351627349854,
    "eval_runtime": 277.9429,
    "eval_samples": 2000,
    "eval_samples_per_second": 7.196,
    "eval_steps_per_second": 0.227,
    "eval_use_label": 3014.730224609375
}