{ "epoch": 2.999098751126561, "eval_kl/chosen": -0.0002014702622545883, "eval_kl/rejected": -0.0002317803300684318, "eval_logits/chosen": -2.109375, "eval_logits/rejected": -2.171875, "eval_logps/chosen": -334.0, "eval_logps/rejected": -322.0, "eval_loss": 0.4495999217033386, "eval_nll_loss": 1.15625, "eval_ref_probs/chosen": 0.0038271306548267603, "eval_ref_probs/rejected": 0.0034569373819977045, "eval_rejected_term/max": -0.0002317803300684318, "eval_rejected_term/min": -0.0002317803300684318, "eval_rewards/accuracies": 0.7720000147819519, "eval_rewards/chosen": -1.28125, "eval_rewards/margins": 1.234375, "eval_rewards/rejected": -2.515625, "eval_runtime": 25.1561, "eval_samples_per_second": 39.752, "eval_steps_per_second": 9.938 }