Scientific QA on MedQA
83.74AccuracyVoting
Evaluation Results
| Method | Links | |
|---|---|---|
| VotingBackbone=Qwen3-14B, Workers=32026.06 | 83.74 | |
| Parallel-SynthesisBackbone=Qwen3-14B, Workers=3, Merging=Weighted averaging (lambda=0.5)2026.06 | 83.58 | |
| SingleBackbone=Qwen3-14B, Workers=12026.06 | 82.72 | |
| Text-SerializationBackbone=Qwen3-14B, Workers=32026.06 | 82.72 | |
| KVLINKBackbone=Qwen3-14B, Workers=32026.06 | 67.05 | |
| CacheBlendBackbone=Qwen3-14B, Workers=32026.06 | 40.38 | |
| APEBackbone=Qwen3-14B, Workers=32026.06 | 17.05 |