Biomedical QA Classification on PubMedQA
94AccuracyHiLight
Evaluation results
| Method | Links | ||
|---|---|---|---|
| HiLightSolver=Qwen3-14B, Actor model=Qwen3-0.6B, Highlight budget (γ)=0.152026.04 | 94 | 82.1 | |
| DSPy (MIPROv2)Solver=Qwen3-14B2026.04 | 93 | 76.7 | |
| APESolver=Qwen3-14B2026.04 | 92.6 | 72.7 | |
| PRLSolver=Qwen3-14B2026.04 | 92.4 | 73.5 | |
| BFRSSolver=Qwen3-14B2026.04 | 92.2 | 76.1 | |
| OPROSolver=Qwen3-14B2026.04 | 92.2 | 76.8 | |
| MISolver=Qwen3-14B2026.04 | 92.1 | 77.6 |