{ "epoch": 2.999098751126561, "eval_kl/chosen": -0.00020001459051854908, "eval_kl/rejected": 0.00019838479056488723, "eval_logits/chosen": -1.4375, "eval_logits/rejected": -1.3984375, "eval_logps/chosen": -322.0, "eval_logps/rejected": -298.0, "eval_loss": 0.6905937790870667, "eval_nll_loss": 1.0546875, "eval_ref_probs/chosen": 0.0038271306548267603, "eval_ref_probs/rejected": 0.0034569373819977045, "eval_rejected_term/max": 0.00019838479056488723, "eval_rejected_term/min": 0.00019838479056488723, "eval_rewards/accuracies": 0.5799999833106995, "eval_rewards/chosen": -0.0018310546875, "eval_rewards/margins": 0.10400390625, "eval_rewards/rejected": -0.10595703125, "eval_runtime": 25.2913, "eval_samples_per_second": 39.539, "eval_steps_per_second": 9.885 }