List your company

Mathematical Reasoning on GSM8K (Pass@1, TPR/F1 vs FPR)

38.6Pass@1UW
20.50425.20229.934.598Apr 24, 2026

Evaluation results

MethodLinks
UWModel=LLaMA-3-8B2026.0438.65.31012.921.9
SWEETModel=LLaMA-3-8B2026.0438.679.588.294.795.1
KGWModel=LLaMA-3-8B2026.0437.163.677.465.276.8
EWDModel=LLaMA-3-8B2026.0435.674.284.895.595.5
EWD + SSGModel=LLaMA-3-8B2026.0435.699.299.299.297.4
KGW + SSGModel=LLaMA-3-8B2026.0434.189.4949195.5
SWEET + SSGModel=LLaMA-3-8B2026.0434.110099.610097.8
UWModel=DSMath-7B2026.0427.34.58.66.812.2
SWEETModel=DSMath-7B2026.0427.394.796.99796.2
KGW + SSGModel=DSMath-7B2026.0425.890.994.991.793.4
SWEET + SSGModel=DSMath-7B2026.042594.796.99796.2
EWD + SSGModel=DSMath-7B2026.042596.297.79796.2
KGWModel=DSMath-7B2026.0421.741.758.557.671
EWDModel=DSMath-7B2026.0421.293.296.195.595.5