List your company

Mathematical Reasoning on GSM-8K

95.6AccuracyOriginal
94.5694.8395.195.37Feb 6, 2026

Evaluation results

MethodLinks
OriginalBatch Size=82026.0295.650.711.3
Algorithm 6 (1, 5)Batch Size=8, k0=1, mg=52026.0294.633.25.94