List your company

Science Question Answering on GPQA (Score)

32.84ScoreDPO
-1.0517287.74708616.545925.344714May 6, 2026May 9, 2026May 12, 2026May 16, 2026May 19, 2026May 22, 2026May 26, 2026

Evaluation results

MethodLinks
DPOIteration=12026.0532.84
DEPOIteration=12026.0532.83
DEPOIteration=32026.0532.5
XPOIteration=12026.0532.47
Llama-3-8B-SFT2026.0532.4
POPOIteration=12026.0532.38
DEPOIteration=22026.0532.19
POPOIteration=32026.0531.94
POPOIteration=22026.0531.72
XPOIteration=22026.0531.25
DPOIteration=32026.0531.24
XPOIteration=32026.0531.01
DPOIteration=22026.0530.76
AdamWModel=MOE-68B-A3B, Training Tokens=700B, Evaluation Mode=FewShot2026.050.2743
MONAModel=MOE-68B-A3B, Training Tokens=700B, Evaluation Mode=FewShot2026.050.2555
MuonModel=MOE-68B-A3B, Training Tokens=700B, Evaluation Mode=FewShot2026.050.2518