Science Question Answering on GPQA (Score)
32.84ScoreDPO
Evaluation Results
| Method | Links | |
|---|---|---|
| DPOIteration=12026.05 | 32.84 | |
| DEPOIteration=12026.05 | 32.83 | |
| DEPOIteration=32026.05 | 32.5 | |
| XPOIteration=12026.05 | 32.47 | |
| Llama-3-8B-SFT2026.05 | 32.4 | |
| POPOIteration=12026.05 | 32.38 | |
| DEPOIteration=22026.05 | 32.19 | |
| POPOIteration=32026.05 | 31.94 | |
| POPOIteration=22026.05 | 31.72 | |
| XPOIteration=22026.05 | 31.25 | |
| DPOIteration=32026.05 | 31.24 | |
| XPOIteration=32026.05 | 31.01 | |
| DPOIteration=22026.05 | 30.76 | |
| AdamWModel=MOE-68B-A3B, Training Tokens=700B, Evaluation Mode=FewShot2026.05 | 0.2743 | |
| MONAModel=MOE-68B-A3B, Training Tokens=700B, Evaluation Mode=FewShot2026.05 | 0.2555 | |
| MuonModel=MOE-68B-A3B, Training Tokens=700B, Evaluation Mode=FewShot2026.05 | 0.2518 |