Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

MAQA

Benchmarks

Task NameDataset NameSOTA ResultTrend
Multi-answer Question AnsweringMAQA-ΔK−1
KL Divergence-0.149
48
Uncertainty QuantificationMAQA ∆K−1
KL Divergence AUC0.757
28
Uncertainty QuantificationMAQA
Hamming AUC83.5
28
Multi-answer Question AnsweringMAQA
Hamming Distance0.04
28
Multi-answer Question Answering (Sets)MAQA {0, 1}^K
Hamming Score102.3
20
Question AnsweringMAQA
Accuracy0.635
7
Error DetectionMAQA* High-Ambiguity Subset H[p*] >= 1.5 (test)
AUROC65
5
ClassificationMAQA (test)
Accuracy63.5
5
Showing 8 of 8 rows