Biomedical Question Answering on PubMedQA
0.6Baseline AccuracyThinkSwitch
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| ThinkSwitchModel Variant=thinking2026.05 | 0.6 | 0.8333 | 0.8333 | 0.2333 | |
| ThinkSwitchModel Variant=instruct2026.05 | 0.4333 | 0.6333 | 0.6 | 0.1667 |
Loading the SOTA2 catalog…