試験
通常
試験時間: 00:00問題時間: 00:00
問題一覧
問題 63 / 65
Which method generates multiple response candidates for the same prompt, ranks them using automatic evaluation metrics or scores, and selects the best response?
①Utilizing Chain of Thought
②How to measure robustness using adversarial inputs
③How to review everything manually
④Ranking or rescoring of n-best candidates
⑤Token-level output constraints