Multi-task Language Understanding on MMLU (Pass@1)
64.47MMLU Pass@1 AccuracyHEX
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| HEXGen. Len.=128, Training Necessity=Training-free2026.02 | 64.47 | 1,600 | |
| DVOTINGGen. Len.=128, Training Necessity=Training-free2026.02 | 63.1 | 224 | |
| Majority VotingGen. Len.=128, Training Necessity=Training-free2026.02 | 62.91 | 320 | |
| LLaDA-1.5*Gen. Len.=128, Training Necessity=None2026.02 | 58.27 | 128 |