List your company

Arithmetic Reasoning on MNAdd 5dgt

57.79AccuracyVLC
-0.314814.770129.85544.9399Mar 25, 2026

Evaluation results

MethodLinks
VLCBackbone Scale=7B, Random Seeds=52026.0357.79
PrismBackbone Scale=7B, Random Seeds=52026.0346.21
End2end RSBackbone Scale=7B, Random Seeds=52026.036.95
End2end FTBackbone Scale=7B, Random Seeds=52026.032.53
ViperGPTBackbone Scale=7B, Random Seeds=52026.031.92