Multi-turn information acquisition on Beauty
82.9Success RateCUP
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| CUPModel=Mistral-7B-v0.32026.04 | 82.9 | 4.06 | |
| CUPModel=Qwen3-4B2026.04 | 82.12 | 4.1 | |
| CUPModel=Llama-3.1-8B2026.04 | 81.99 | 4.06 | |
| MISQ-HFModel=Llama-3.1-8B2026.04 | 69.95 | 4.31 | |
| BED-LLMModel=Llama-3.1-8B2026.04 | 68.78 | 4.38 | |
| MISQ-HFModel=Mistral-7B-v0.32026.04 | 64.9 | 4.3 | |
| MISQ-HFModel=Qwen3-4B2026.04 | 63.21 | 4.33 | |
| BED-LLMModel=Mistral-7B-v0.32026.04 | 61.53 | 4.62 | |
| UoTModel=Llama-3.1-8B2026.04 | 57.9 | 4.69 | |
| BED-LLMModel=Qwen3-4B2026.04 | 55.57 | 4.49 | |
| UoTModel=Mistral-7B-v0.32026.04 | 53.11 | 4.73 | |
| UoTModel=Qwen3-4B2026.04 | 51.68 | 4.67 | |
| ATDModel=Mistral-7B-v0.32026.04 | 49.22 | 4.69 | |
| ATDModel=Llama-3.1-8B2026.04 | 49.09 | 4.66 | |
| ATDModel=Qwen3-4B2026.04 | 48.96 | 4.7 | |
| LLM-planningModel=Llama-3.1-8B2026.04 | 44.43 | 4.75 | |
| LLM-planningModel=Qwen3-4B2026.04 | 44.3 | 4.72 | |
| LLM-planningModel=Mistral-7B-v0.32026.04 | 43.01 | 4.81 | |
| CoTModel=Qwen3-4B2026.04 | 38.47 | 4.11 | |
| CoTModel=Llama-3.1-8B2026.04 | 37.82 | 4.1 | |
| Direct PromptingModel=Qwen3-4B2026.04 | 37.05 | 4.18 | |
| CoTModel=Mistral-7B-v0.32026.04 | 35.42 | 4.17 | |
| Direct PromptingModel=Llama-3.1-8B2026.04 | 33.94 | 4.22 | |
| Direct PromptingModel=Mistral-7B-v0.32026.04 | 32.25 | 4.2 | |
| SBERT2026.04 | 11.14 | 4.82 |