Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

AdvGLUE

Benchmarks

Task NameDataset NameSOTA ResultTrend
HelpfulnessAdvGLUE
Accuracy75.15
20
Binary classificationAdvGLUE (test)
QNLI Accuracy0.701
17
Natural Language UnderstandingAdvGLUE
RTE Accuracy67.9
8
Adversarial RobustnessAdvGLUE (test)
AdvGLUE Score76.73
6
Showing 4 of 4 rows