{ "epoch": 3.0, "eval_logits/chosen": -4.191605567932129, "eval_logits/rejected": -4.081777572631836, "eval_logps/chosen": -402.4017333984375, "eval_logps/rejected": -315.6105651855469, "eval_loss": 0.6876626014709473, "eval_rewards/accuracies": 0.5644999742507935, "eval_rewards/chosen": 0.025381002575159073, "eval_rewards/margins": 0.011920945718884468, "eval_rewards/rejected": 0.013460054062306881, "eval_runtime": 777.3765, "eval_samples": 2000, "eval_samples_per_second": 2.573, "eval_steps_per_second": 0.643 }