List your company

Logical Reasoning on KandLogic 5obj

91.18AccuracyViperGPT
53.615263.367673.1282.8724Mar 25, 2026

Evaluation results

MethodLinks
ViperGPTBackbone Scale=7B, Random Seeds=52026.0391.18
VLCBackbone Scale=7B, Random Seeds=52026.0389.97
End2end FTBackbone Scale=7B, Random Seeds=52026.0364.28
End2end RSBackbone Scale=7B, Random Seeds=52026.0362.11
PrismBackbone Scale=7B, Random Seeds=52026.0355.06