Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Automated Probing on SuperGLUE
Loading...
38
Error Rate
AutoDetect
36.68
45.59
54.5
63.41
Feb 13, 2026
Error Rate
Attack Success Rate
Updated 1mo ago
Evaluation Results
Method
Method
Links
Error Rate
Attack Success Rate
AutoDetect
Generator Model=GPT-5....
2026.02
38
-
PAIR
Generator Model=GPT-5....
2026.02
44
89.5
PROBELLM
Generator Model=GPT-5....
2026.02
71
-
Feedback
Search any
task
Search any
task
