List your company

Mathematical Reasoning on GSM-Hard GSM8k-Aug rationale format (original)

13.04Pass@1SFT-CoT
5.1887.22659.26511.3035May 25, 2026

Evaluation results

MethodLinks
SFT-CoTRationale Format=GSM8k-Aug2026.0513.0434.6
Ours-RLRationale Format=GSM8k-Aug2026.0510.7721.03
Ours-SFTRationale Format=GSM8k-Aug2026.059.3322.35
CoLaR-2Rationale Format=GSM8k-Aug2026.059.0814
CoconutRationale Format=GSM8k-Aug2026.055.496