{ "epoch": 2.999098751126561, "eval_kl/chosen": -0.00013887668319512159, "eval_kl/rejected": -4.645527224056423e-05, "eval_logits/chosen": -2.28125, "eval_logits/rejected": -2.359375, "eval_logps/chosen": -340.0, "eval_logps/rejected": -332.0, "eval_loss": 0.416765034198761, "eval_nll_loss": 1.234375, "eval_ref_probs/chosen": 0.0038271306548267603, "eval_ref_probs/rejected": 0.0034569373819977045, "eval_rejected_term/max": -4.645527224056423e-05, "eval_rejected_term/min": -4.645527224056423e-05, "eval_rewards/accuracies": 0.7839999794960022, "eval_rewards/chosen": -1.84375, "eval_rewards/margins": 1.578125, "eval_rewards/rejected": -3.421875, "eval_runtime": 25.2159, "eval_samples_per_second": 39.658, "eval_steps_per_second": 9.914 }