{
    "epoch": 2.999098751126561,
    "eval_kl/chosen": 0.00019685739243868738,
    "eval_kl/rejected": 0.00023367426183540374,
    "eval_logits/chosen": -1.3671875,
    "eval_logits/rejected": -1.328125,
    "eval_logps/chosen": -322.0,
    "eval_logps/rejected": -298.0,
    "eval_loss": 0.8147246241569519,
    "eval_nll_loss": 1.046875,
    "eval_ref_probs/chosen": 0.0038271306548267603,
    "eval_ref_probs/rejected": 0.0034569373819977045,
    "eval_rejected_term/max": 0.00023367426183540374,
    "eval_rejected_term/min": 0.00023367426183540374,
    "eval_rewards/accuracies": 0.4880000054836273,
    "eval_rewards/chosen": 0.00518798828125,
    "eval_rewards/margins": 0.04833984375,
    "eval_rewards/rejected": -0.043212890625,
    "eval_runtime": 25.4589,
    "eval_samples_per_second": 39.279,
    "eval_steps_per_second": 9.82
}