List your company

Knowledge-intensive Multiple-Choice QA on MMLU

1Spearman RhoSEAL
0.69840.77670.8550.9333May 28, 2026

Evaluation results

MethodLinks
SEALProtocol=SEAL2026.0511.3
FullPairProtocol=FullPair2026.0511.77
FixedRubProtocol=FixedRub2026.050.950.77
FlatBrkProtocol=FlatBrk2026.050.931.3
PointwiseProtocol=Pointwise2026.050.861.13
OriginalProtocol=Original2026.050.791
ListwiseProtocol=Listwise2026.050.713.52