Persuasion on Cornell CMV Persuadee: Phi-4 OOD
33.09Agreement ShiftGemma-3
Evaluation Results
| Method | Links | |
|---|---|---|
| Gemma-3Size=27B2025.05 | 33.09 | |
| DS-R1Size=671B2025.05 | 32.31 | |
| LLaMa3.2 + ToMAPSize=3B2025.05 | 31.64 | |
| LLaMa3.2 + RLSize=3B2025.05 | 25.69 | |
| GPT-4oSize=N/A2025.05 | 23.97 | |
| Qwen2.5 + ToMAPSize=3B2025.05 | 20.93 | |
| LLaMa3.1Size=70B2025.05 | 19.84 | |
| Qwen2.5Size=3B2025.05 | 19.44 | |
| LLaMa3.2Size=3B2025.05 | 18.86 | |
| Qwen2.5 + SFTSize=3B2025.05 | 17.95 | |
| LLaMa3.2 + SFTSize=3B2025.05 | 16.41 | |
| Qwen2.5 + RLSize=3B2025.05 | 16.28 |