Multiple-choice Question Answering on PubMedQA (Accuracy)
63.62AccuracyDS2-INSTRUCT
Evaluation Results
| Method | Links | |
|---|---|---|
| DS2-INSTRUCTModel Family=Qwen2.52026.03 | 63.62 | |
| InstructMixModel Family=Qwen2.52026.03 | 62.15 | |
| Self-InstructModel Family=Qwen2.52026.03 | 61.47 | |
| ExploreInstructModel Family=Qwen2.52026.03 | 60.89 | |
| Zero-ShotModel Family=Qwen2.52026.03 | 60.34 | |
| DS2-INSTRUCTModel Family=Llama32026.03 | 48.56 | |
| DS2-INSTRUCTModel Family=Mistral2026.03 | 39.11 | |
| InstructMixModel Family=Mistral2026.03 | 31.48 | |
| ExploreInstructModel Family=Mistral2026.03 | 28.35 | |
| InstructMixModel Family=Llama32026.03 | 27.41 | |
| Self-InstructModel Family=Mistral2026.03 | 26.83 | |
| Self-InstructModel Family=Llama32026.03 | 25.34 | |
| Zero-ShotModel Family=Llama32026.03 | 24.92 | |
| ExploreInstructModel Family=Llama32026.03 | 24.77 | |
| Zero-ShotModel Family=Mistral2026.03 | 14.92 |