List your company

Arithmetic Reasoning on MNAdd 3dgt

79.65AccuracyEnd2end FT
17.312433.496249.6865.8638Mar 25, 2026

Evaluation results

MethodLinks
End2end FTBackbone Scale=7B, Random Seeds=52026.0379.65
VLCBackbone Scale=7B, Random Seeds=52026.0354.62
PrismBackbone Scale=7B, Random Seeds=52026.0350.22
End2end RSBackbone Scale=7B, Random Seeds=52026.0326.99
ViperGPTBackbone Scale=7B, Random Seeds=52026.0319.71