List your company

Mathematical Reasoning on GSM8K (GSM8K Metric)

51.1GSM8K AccuracyTA w/ WSC
38.20441.55244.948.248Dec 22, 2025

Evaluation results

MethodLinks
TA w/ WSCBackbone=Llama-2-7b2025.1251.1
TSV-MBackbone=Llama-2-7b2025.1249.7
DAREBackbone=Llama-2-7b2025.1247
TIESBackbone=Llama-2-7b2025.1245
TABackbone=Llama-2-7b2025.1243.8
Iso-CBackbone=Llama-2-7b2025.1242
Iso-CTSBackbone=Llama-2-7b2025.1238.7