List your company

Question Answering on StrategyQA (Pearson r)

0.94Pearson rOC
-0.01680.23160.480.7284Jun 11, 2026

Evaluation results

MethodLinks
OCFormat=binarized at 0.5, Signal=paper’s primary2026.060.94
OCFormat=continuous2026.060.94
−HK=10, Format=continuous2026.060.8
naive direct-answer SCK=10, Mechanism=ablation2026.060.7
naive direct-answer SCK=3, Mechanism=ablation2026.060.67
decomposed-SCK=102026.060.47
CoT-SCK=10, Format=continuous2026.060.46
CoT-SCK=3, Format=continuous2026.060.45
RLHFlow-PRMMechanism=min over steps2026.060.28
P(True)Format=continuous, Models=9 models2026.060.04
Skywork-PRMMechanism=min over steps2026.060.02