List your company

Mathematical Reasoning on GSM8K (test) (Accuracy, CR, OR)

23.19AccuracyDGAO
15.18217.26119.3421.419May 12, 2026

Evaluation results

MethodLinks
DGAOBackbone=Llama-3.2-3B2026.0523.1958.4734.72
PromptBackbone=Llama-3.2-3B2026.0522.9956.5937.81
DEmOBackbone=Llama-3.2-3B2026.0522.9556.9137.45
GlobalEBackbone=Llama-3.2-3B2026.0522.8956.8537.51
SFTBackbone=Llama-3.2-3B2026.0522.8556.7837.56
SPHSBackbone=Llama-3.2-3B2026.0522.4856.1237.38
PINEBackbone=Llama-3.2-3B2026.0515.4949.8838.19