{
    "epoch": 2.999098751126561,
    "eval_kl/chosen": 0.00031474779825657606,
    "eval_kl/rejected": 0.0005000739474780858,
    "eval_logits/chosen": -1.4140625,
    "eval_logits/rejected": -1.3828125,
    "eval_logps/chosen": -322.0,
    "eval_logps/rejected": -298.0,
    "eval_loss": 0.712820291519165,
    "eval_nll_loss": 1.0546875,
    "eval_ref_probs/chosen": 0.0038271306548267603,
    "eval_ref_probs/rejected": 0.0034569373819977045,
    "eval_rejected_term/max": 0.0005000739474780858,
    "eval_rejected_term/min": 0.0005000739474780858,
    "eval_rewards/accuracies": 0.5640000104904175,
    "eval_rewards/chosen": -0.052001953125,
    "eval_rewards/margins": 0.11083984375,
    "eval_rewards/rejected": -0.1630859375,
    "eval_runtime": 25.3354,
    "eval_samples_per_second": 39.47,
    "eval_steps_per_second": 9.868
}