Multi-turn dialogue on MedDG
86.77Success Rate (SR)MCTS Router
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| MCTS RouterCandidate LLM Set=Qwen Series Set2026.04 | 86.77 | 4.77 | |
| MCTS RouterCandidate LLM Set=Llama Series Set2026.04 | 82.38 | 6.41 | |
| Greedy RouterCandidate LLM Set=Qwen Series Set2026.04 | 79.71 | 4.95 | |
| DialRouterCandidate LLM Set=Qwen Series Set2026.04 | 79.03 | 4.98 | |
| Prompt LLMCandidate LLM Set=Qwen Series Set2026.04 | 72.74 | 5.29 | |
| Greedy RouterCandidate LLM Set=Llama Series Set2026.04 | 72.51 | 6.3 | |
| Qwen3-14BCandidate LLM Set=Qwen Series Set2026.04 | 72.5 | 5.27 | |
| RouterDCCandidate LLM Set=Qwen Series Set2026.04 | 72.5 | 5.27 | |
| MF RouterCandidate LLM Set=Qwen Series Set2026.04 | 72.41 | 4.97 | |
| KNN RouterCandidate LLM Set=Qwen Series Set2026.04 | 71.99 | 4.75 | |
| Qwen3-8BCandidate LLM Set=Qwen Series Set2026.04 | 71.82 | 5.02 | |
| AvengersCandidate LLM Set=Qwen Series Set2026.04 | 71.69 | 5 | |
| RandomCandidate LLM Set=Qwen Series Set2026.04 | 71.46 | 4.86 | |
| Qwen3-4BCandidate LLM Set=Qwen Series Set2026.04 | 71.04 | 5.03 | |
| DialRouterCandidate LLM Set=Llama Series Set2026.04 | 70.65 | 6.16 | |
| MF RouterCandidate LLM Set=Llama Series Set2026.04 | 66.19 | 6.3 | |
| AvengersCandidate LLM Set=Llama Series Set2026.04 | 66.14 | 6.65 | |
| RouterDCCandidate LLM Set=Llama Series Set2026.04 | 66.09 | 6.36 | |
| Llama3.1-8B-InstructCandidate LLM Set=Llama Series Set2026.04 | 65.35 | 6.19 | |
| Prompt LLMCandidate LLM Set=Llama Series Set2026.04 | 65.35 | 6.19 | |
| KNN RouterCandidate LLM Set=Llama Series Set2026.04 | 64.58 | 6.05 | |
| RandomCandidate LLM Set=Llama Series Set2026.04 | 61.11 | 6.34 | |
| Llama3.2-3B-InstructCandidate LLM Set=Llama Series Set2026.04 | 60.78 | 6.44 | |
| Llama3.2-1B-InstructCandidate LLM Set=Llama Series Set2026.04 | 43.61 | 6.97 |