試験
通常
試験時間: 00:00問題時間: 00:00
問題一覧
問題 42 / 65
I want to reduce the risk of generative AI producing misinformation (hallucinations) and improve the accuracy of medical documents. Which approach is most effective?
①Adopt a larger general-purpose model as-is.
②Evaluate the output only through user testing.
③Apply a policy that simply deletes the output.
④Address it only by the translator's manual correction of all output.
⑤Introduction of custom validation rules and a post-output verification pipeline