Question Answering on WMDP multiple-choice QA
65.1Bio AccuracyBase
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| BaseModel=Mistral-7B, Zero-shot=true2026.03 | 65.1 | 41.5 | |
| BaseModel=Zephyr-7B-β, Zero-shot=true2026.03 | 64.3 | 44.8 | |
| LLMUModel=Zephyr-7B-β, Zero-shot=true2026.03 | 59.5 | 39.5 | |
| SSDModel=Zephyr-7B-β, Zero-shot=true2026.03 | 50.2 | 35 | |
| SCRUBModel=Zephyr-7B-β, Zero-shot=true2026.03 | 43.8 | 39.3 | |
| ASUModel=Zephyr-7B-β, Zero-shot=true2026.03 | 32.1 | 31.7 | |
| ASUModel=Mistral-7B, Zero-shot=true2026.03 | 31.5 | 29.5 | |
| RMUModel=Zephyr-7B-β, Zero-shot=true2026.03 | 31.2 | 28.2 | |
| RMUModel=Mistral-7B, Zero-shot=true2026.03 | 30.7 | 32.3 |