{ "epoch": 2.999098751126561, "eval_kl/chosen": 0.0002554964448790997, "eval_kl/rejected": -2.5187298888340592e-05, "eval_logits/chosen": -1.625, "eval_logits/rejected": -1.5703125, "eval_logps/chosen": -290.0, "eval_logps/rejected": -310.0, "eval_loss": 0.693246066570282, "eval_nll_loss": 0.94140625, "eval_ref_probs/chosen": 0.004795589949935675, "eval_ref_probs/rejected": 0.004843756556510925, "eval_rejected_term/max": -2.5187298888340592e-05, "eval_rejected_term/min": -2.5187298888340592e-05, "eval_rewards/accuracies": 0.7039999961853027, "eval_rewards/chosen": 0.0027008056640625, "eval_rewards/margins": 0.220703125, "eval_rewards/rejected": -0.21875, "eval_runtime": 25.299, "eval_samples_per_second": 39.527, "eval_steps_per_second": 9.882 }