試験
通常
試験時間: 00:00問題時間: 00:00
問題一覧
問題 16 / 65
Which evaluation method is most reliable when comparing multiple prompt candidates to choose the best one?
①Automatic comparison of generated token lengths and character counts
②Comparison of average model inference times
③Human evaluation of output quality
④Comparison of model parameter counts and sizes
⑤Choose randomly and use.