Negotiation on Hospital (test)
100Success RateVanilla (LLM)
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Vanilla (LLM)Backbone=Qwen3.5-Plus2026.05 | 100 | 35 | 4.5 | |
| Random2026.05 | 100 | 30 | 4.2 | |
| IQLBackbone=Qwen3.5-Plus2026.05 | 100 | 0 | 3.1 | |
| IQL+JPOBackbone=Qwen2.5-7B-Instruct2026.05 | 100 | 35 | 7.4 | |
| IQL+SFT+JPOBackbone=Qwen2.5-7B-Instruct, Note=Full EmoDistill pipeline2026.05 | 100 | 45 | 5.5 | |
| IQL+SFTBackbone=Qwen2.5-7B-Instruct2026.05 | 95 | 20 | 6 | |
| Vanilla (SLM)Backbone=Qwen2.5-7B-Instruct2026.05 | 90 | 40.3 | 4.3 |