WTQ
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
WTQ Mix ROBUT (test)
64.5Accuracy (Pre-perturbation)
5
WTQ Column Masking ROBUT (test)
60.4Accuracy (Pre-Perturbation)
5
ROBUT-WTQ (dev)
61Accuracy
5
WTQ I (held-in)
73TEDS Score
3
WTQ
14.71FE
3