{ "epoch": 2.888888888888889, "eval_logits/chosen": -2.356201410293579, "eval_logits/rejected": -2.384758234024048, "eval_logps/chosen": -262.8277282714844, "eval_logps/rejected": -213.3736114501953, "eval_loss": 0.6153290867805481, "eval_rewards/accuracies": 0.875, "eval_rewards/chosen": 1.9784826040267944, "eval_rewards/margins": 1.742444634437561, "eval_rewards/rejected": 0.2360377162694931, "eval_runtime": 24.7523, "eval_samples_per_second": 15.352, "eval_steps_per_second": 0.242 }