List your company

Arithmetic Reasoning on MNAdd 7dgt

52.06AccuracyVLC
-2.040812.004626.0540.0954Mar 25, 2026

Evaluation results

MethodLinks
VLCBackbone Scale=7B, Random Seeds=52026.0352.06
PrismBackbone Scale=7B, Random Seeds=52026.0330.69
End2end RSBackbone Scale=7B, Random Seeds=52026.031.85
End2end FTBackbone Scale=7B, Random Seeds=52026.031.35
ViperGPTBackbone Scale=7B, Random Seeds=52026.030.04