{
    "epoch": 2.999098751126561,
    "eval_kl/chosen": 0.003269560867920518,
    "eval_kl/rejected": 0.003505800385028124,
    "eval_logits/chosen": -1.7265625,
    "eval_logits/rejected": -1.703125,
    "eval_logps/chosen": -324.0,
    "eval_logps/rejected": -304.0,
    "eval_loss": 1.6438730955123901,
    "eval_nll_loss": 1.0546875,
    "eval_ref_probs/chosen": 0.0038271306548267603,
    "eval_ref_probs/rejected": 0.0034569373819977045,
    "eval_rejected_term/max": 0.003505800385028124,
    "eval_rejected_term/min": 0.003505800385028124,
    "eval_rewards/accuracies": 0.7120000123977661,
    "eval_rewards/chosen": -0.267578125,
    "eval_rewards/margins": 0.39453125,
    "eval_rewards/rejected": -0.6640625,
    "eval_runtime": 26.0385,
    "eval_samples_per_second": 38.405,
    "eval_steps_per_second": 9.601
}