List your company

Mistake Correction on BIG-Bench Mistake (BBM)

15Initial AccuracyDISC
14.2514.6251515.375Jun 19, 2026

Evaluation results

MethodLinks
DISCModel=Claude Sonnet 4.5, K=32026.061581.61,195814981.3
CoVeModel=Claude Sonnet 4.52026.061558.68517511.358
Self-RefineModel=Claude Sonnet 4.52026.061549.66332130.142
DISCModel=GPT-4o, K=32026.061549.9666183743.7
CoVeModel=GPT-4o2026.0615629164520.460.2
Self-RefineModel=GPT-4o2026.061527.6289575.118.5