{
    "epoch": 2.999098751126561,
    "eval_kl/chosen": 0.0011769378324970603,
    "eval_kl/rejected": -0.0004123514227103442,
    "eval_logits/chosen": -1.1875,
    "eval_logits/rejected": -1.1640625,
    "eval_logps/chosen": -328.0,
    "eval_logps/rejected": -300.0,
    "eval_loss": 1.0556386709213257,
    "eval_nll_loss": 1.0703125,
    "eval_ref_probs/chosen": 0.0038271306548267603,
    "eval_ref_probs/rejected": 0.0034569373819977045,
    "eval_rejected_term/max": -0.0004123514227103442,
    "eval_rejected_term/min": -0.0004123514227103442,
    "eval_rewards/accuracies": 0.35199999809265137,
    "eval_rewards/chosen": -0.59765625,
    "eval_rewards/margins": -0.259765625,
    "eval_rewards/rejected": -0.337890625,
    "eval_runtime": 25.3972,
    "eval_samples_per_second": 39.374,
    "eval_steps_per_second": 9.844
}