Persuasion on Cornell CMV, Anthropic and args.me Aggregate
24.35Avg. Agreement Shift (%)LLaMa3.2 + ToMAP
Evaluation Results
| Method | Links | |
|---|---|---|
| LLaMa3.2 + ToMAPSize=3B2025.05 | 24.35 | |
| LLaMa3.2 + RLSize=3B2025.05 | 21.76 | |
| Qwen2.5 + ToMAPSize=3B2025.05 | 17.48 | |
| Gemma-3Size=27B2025.05 | 17.27 | |
| DS-R1Size=671B2025.05 | 17.02 | |
| Qwen2.5 + RLSize=3B2025.05 | 14.6 | |
| LLaMa3.2 + SFTSize=3B2025.05 | 12.68 | |
| GPT-4oSize=N/A2025.05 | 12.54 | |
| Qwen2.5 + SFTSize=3B2025.05 | 12.49 | |
| LLaMa3.2Size=3B2025.05 | 12.37 | |
| Qwen2.5Size=3B2025.05 | 11.76 | |
| LLaMa3.1Size=70B2025.05 | 10.57 |