Response Generation on RolePlayConv (evaluation set)
7.4Response ScoreModeratorLM-Think
Evaluation Results
| Method | Links | |
|---|---|---|
| ModeratorLM-ThinkEvaluation Method=LLM-as-a-judge2026.06 | 7.4 | |
| ModeratorLMEvaluation Method=LLM-as-a-judge2026.06 | 6.9 | |
| MP-BaselineEvaluation Method=LLM-as-a-judge2026.06 | 4.6 |