List your company

Mathematical Reasoning on Average (GSM8K, MATH, MMLU-MATH, MATHQA) (test)

51.4AccuracyDR+DA
31.53636.69341.8547.007Jan 29, 2026

Evaluation results

MethodLinks
DR+DADepth=32, 0-SHOT=true2026.0151.43.5
DRDepth=32, 0-SHOT=true2026.0147.62.6
DRDepth=16, 0-SHOT=true2026.0145.74
DR+DADepth=16, 0-SHOT=true2026.0143.42.9
LADepth=32, 0-SHOT=true2026.0139.51
LADepth=16, 0-SHOT=true2026.0132.31