List your company

Mathematical Question Answering on GSM8K (test)

1.16RAvgCluCERT
0.78560.88280.981.0772Dec 1, 2025

Evaluation results

MethodLinks
CluCERTBase Model=ChatGPT-3.5, Setting=Black-box2025.121.161.07
CluCERT(-Clu)Base Model=ChatGPT-3.5, Setting=Black-box, Clustering=Disabled2025.121.051.15
SelfDenoiseBase Model=ChatGPT-3.5, Setting=Black-box, Module=Refine2025.120.921.25
RanMASKBase Model=ChatGPT-3.5, Setting=Black-box, Module=Refine2025.120.81.35