List your company

Question Answering on MuSiQue (Pearson r)

0.96Pearson rOC
0.14880.35940.570.7806Jun 11, 2026

Evaluation results

MethodLinks
OCFormat=continuous2026.060.96
OCFormat=binarized at 0.5, Signal=paper’s primary2026.060.94
CoT-SCK=3, Format=continuous2026.060.45
naive direct-answer SCK=10, Mechanism=ablation2026.060.45
naive direct-answer SCK=3, Mechanism=ablation2026.060.45
−HK=10, Format=continuous2026.060.44
CoT-SCK=10, Format=continuous2026.060.43
P(True)Format=continuous, Models=9 models2026.060.42
Skywork-PRMMechanism=min over steps2026.060.38
decomposed-SCK=102026.060.23
RLHFlow-PRMMechanism=min over steps2026.060.18