General Knowledge on GPQA-D (test)
36.6Score (%)Dream-7B-Base
Evaluation results
| Method | Links | |
|---|---|---|
| Dream-7B-BaseNumber of shots=52026.06 | 36.6 | |
| FReDA-4BNumber of shots=5, Evaluation protocol=Self-refine2026.06 | 36.27 | |
| FReDA-4BNumber of shots=5, Evaluation protocol=Best-of-N2026.06 | 36.27 | |
| Qwen3-4B-BaseNumber of shots=52026.06 | 33.68 | |
| BlockDiff-4BNumber of shots=52026.06 | 33.68 | |
| LLaDA-MoE-7B-A1B-BaseNumber of shots=52026.06 | 28.5 | |
| Qwen2.5-3B-BaseNumber of shots=52026.06 | 26.3 | |
| LLaDA-8B-BaseNumber of shots=52026.06 | 25.2 | |
| DiffuLLaMA-7BNumber of shots=52026.06 | 22.8 |