List your company
ResearchFormerly WizWand

Performance Prediction on MMLU Pro STEM

0.897Train R-squaredRphi
0.13780.33490.5320.7291Sep 25, 2025
Updated 4mo ago

Evaluation results

MethodLinks
RphiScaling Scenario=1B -> 13B, Phase Transition=Pre-to-Pre2025.090.8970.575
RBRIDGEScaling Scenario=1B -> 13B, Phase Transition=Pre-to-Pre2025.090.8970.574
NLLScaling Scenario=1B -> 13B, Phase Transition=Pre-to-Pre2025.090.20712.582
MPCAScaling Scenario=1B -> 13B, Phase Transition=Pre-to-Pre2025.090.21.495
TEDScaling Scenario=1B -> 13B, Phase Transition=Pre-to-Pre2025.090.1991.572
Acc./p@1Scaling Scenario=1B -> 13B, Phase Transition=Pre-to-Pre2025.090.1671.624