Reasoning on GPQA (test)
0.279AccuracyLLaDA
Evaluation Results
| Method | Links | |
|---|---|---|
| LLaDASFT=false, Sequence Length=256, 0-shot=true2025.11 | 0.279 | |
| EDITSequence Length=256, 0-shot=true, SFT=true2025.11 | 0.277 | |
| LLaDASFT=true, Sequence Length=512, 0-shot=true2025.11 | 0.263 | |
| EDITSequence Length=512, 0-shot=true, SFT=true2025.11 | 0.261 | |
| LLaDASFT=false, Sequence Length=512, 0-shot=true2025.11 | 0.257 | |
| EDITSequence Length=128, 0-shot=true, SFT=true2025.11 | 0.255 | |
| LLaDASFT=true, Sequence Length=128, 0-shot=true2025.11 | 0.23 | |
| LLaDASFT=false, Sequence Length=128, 0-shot=true2025.11 | 0.219 | |
| LLaDASFT=true, Sequence Length=256, 0-shot=true2025.11 | 0.205 |