Medical Reasoning on CMB
84.05Exact Match (EM)GraphWalker
Evaluation Results
| Method | Links | |
|---|---|---|
| GraphWalkerBackbone LLM=Qwen3-14B2026.04 | 84.05 | |
| InfluenceBackbone LLM=Qwen3-14B2026.04 | 83.5 | |
| Delta-KNNBackbone LLM=Qwen3-14B2026.04 | 83.02 | |
| LMS3Backbone LLM=Qwen3-14B2026.04 | 82.02 | |
| SPELLBackbone LLM=Qwen3-14B2026.04 | 81.98 | |
| RandomBackbone LLM=Qwen3-14B2026.04 | 81.5 | |
| GradSelBackbone LLM=Qwen3-14B2026.04 | 81.48 | |
| Semantic-embBackbone LLM=Qwen3-14B2026.04 | 80.52 | |
| CONEBackbone LLM=Qwen3-14B2026.04 | 80.41 | |
| Zero-shotBackbone LLM=Qwen3-14B2026.04 | 80.21 | |
| IDSBackbone LLM=Qwen3-14B2026.04 | 79.57 | |
| Time-seriesBackbone LLM=Qwen3-14B2026.04 | 78.05 |