{
    "epoch": 2.999098751126561,
    "eval_kl/chosen": 0.00026231163064949214,
    "eval_kl/rejected": 7.576159987365827e-05,
    "eval_logits/chosen": -1.6328125,
    "eval_logits/rejected": -1.578125,
    "eval_logps/chosen": -290.0,
    "eval_logps/rejected": -310.0,
    "eval_loss": 0.6860566139221191,
    "eval_nll_loss": 0.94140625,
    "eval_ref_probs/chosen": 0.004795589949935675,
    "eval_ref_probs/rejected": 0.004843756556510925,
    "eval_rejected_term/max": 7.576159987365827e-05,
    "eval_rejected_term/min": 7.576159987365827e-05,
    "eval_rewards/accuracies": 0.6840000152587891,
    "eval_rewards/chosen": 0.008544921875,
    "eval_rewards/margins": 0.2158203125,
    "eval_rewards/rejected": -0.20703125,
    "eval_runtime": 25.3391,
    "eval_samples_per_second": 39.465,
    "eval_steps_per_second": 9.866
}