Multi-turn role-play on ED
92.1Success Rate (SR)EmoLLM
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| EmoLLMBackbone=Qwen3-8B, Method Category=Ours2026.03 | 92.1 | 1.87 | 4.86 | 4.92 | |
| PERMBackbone=Qwen3-8B, Method Category=RL Methods2026.03 | 91.8 | 2.22 | 4.79 | 4.28 | |
| ICL-AIFBackbone=Qwen3-8B, Method Category=Prompting Methods2026.03 | 91.3 | 2.29 | 4.73 | 4.24 | |
| RLVERBackbone=Qwen3-8B, Method Category=RL Methods2026.03 | 91.1 | 2.38 | 4.74 | 4.1 | |
| ARG-PromptBackbone=Qwen3-8B, Method Category=Prompting Methods2026.03 | 91 | 2.41 | 4.67 | 4.1 | |
| RLFF-ESCBackbone=Qwen3-8B, Method Category=RL Methods2026.03 | 90.8 | 2.49 | 4.71 | 4.02 | |
| Best-of-NBackbone=Qwen3-8B, Method Category=Prompting Methods2026.03 | 90.4 | 2.78 | 4.48 | 3.72 | |
| PPDPBackbone=Qwen3-8B, Method Category=SFT Methods2026.03 | 90.2 | 2.41 | 4.63 | 4.12 | |
| ESCoTBackbone=Qwen3-8B, Method Category=SFT Methods2026.03 | 89.7 | 2.64 | 4.66 | 3.88 | |
| Ask-an-ExpertBackbone=Qwen3-8B, Method Category=Prompting Methods2026.03 | 89.4 | 2.71 | 4.51 | 3.97 | |
| Qwen3-8BBackbone=Qwen3-8B, Method Category=Base Model2026.03 | 88.7 | 3.03 | 4.58 | 3.84 | |
| SFTBackbone=Qwen3-8B, Method Category=SFT Methods2026.03 | 87.3 | 2.14 | 4.39 | 4.07 |