Multi-domain persuasion on CToMPersu (test)
80SuccessMA2P
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| MA2PModel=deepseek-v32026.05 | 80 | 40 | 0.088 | |
| MA2PModel=gpt-4o-mini2026.05 | 79 | 40 | 0.107 | |
| MA2PModel=gemini-2.5-flash2026.05 | 69 | 42.2 | 0.106 | |
| Karande et al.Model=deepseek-v32026.05 | 59 | 51 | 0.124 | |
| Karande et al.Model=gemini-2.5-flash2026.05 | 44 | 68 | 0.14 | |
| Karande et al.Model=gpt-4o-mini2026.05 | 43 | 64 | 0.114 | |
| Furumai et al.Model=gpt-4o-mini2026.05 | 21 | 50 | 0.097 | |
| Furumai et al.Model=deepseek-v32026.05 | 19 | 31.9 | 0.091 | |
| Furumai et al.Model=gemini-2.5-flash2026.05 | 18 | 45 | 0.093 |