Personalized response selection on PCogAlignBench LS1->LS1
4.161P ScorePCogAlign
Evaluation results
| Method | Links | ||
|---|---|---|---|
| PCogAlignStrategy=Full2025.06 | 4.161 | 53.3 | |
| PCogAlign (S)Strategy=SFT2025.06 | 4.13 | 51.5 | |
| Self-Refine (S)Strategy=SFT2025.06 | 4.091 | 50.3 | |
| PCogAlign (P)Strategy=Prompt2025.06 | 4.07 | 47.5 | |
| RLAIF (S)Strategy=SFT2025.06 | 4.046 | 45.1 | |
| RAGStrategy=Prompt2025.06 | 4.022 | 43.9 | |
| PCogAlign (D)Strategy=DPO2025.06 | 4.02 | 45.2 | |
| RLCD (S)Strategy=SFT2025.06 | 3.996 | 43.2 | |
| RS PromptStrategy=Prompt2025.06 | 3.989 | 44.1 | |
| RLCD (D)Strategy=DPO2025.06 | 3.985 | 44.6 | |
| RLAIF (D)Strategy=DPO2025.06 | 3.984 | 43.6 | |
| Self-Refine (D)Strategy=DPO2025.06 | 3.982 | 43.7 | |
| RS Prompt (S)Strategy=SFT2025.06 | 3.857 | 35.3 | |
| BaseStrategy=Prompt2025.06 | 3.781 | 0 |