Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Adversarial Attack Detection on Balanced 500-sample non-adaptive
Loading...
94
Precision
SpectralGuard (Multi-Layer)
4.56
27.78
51
74.22
Mar 12, 2026
Precision
Recall
F1 Score
AUC
False Positive Rate
Updated 1mo ago
Evaluation Results
Method
Method
Links
Precision
Recall
F1 Score
AUC
False Positive Rate
SpectralGuard (Multi-Layer)
Setting=Non-adaptive,...
2026.03
94
98
96.1
0.989
6
Internal L2-Norm Monitor
Setting=Non-adaptive,...
2026.03
62
51
56
-
5
SpectralGuard (Single-Layer)
Setting=Non-adaptive,...
2026.03
54.9
70
61.9
-
67.2
Perplexity Filter
Setting=Non-adaptive,...
2026.03
42
31
36
-
2
Pattern Matcher
Setting=Non-adaptive,...
2026.03
15
12
13
-
-
Toxicity Classifier
Setting=Non-adaptive,...
2026.03
8
5
6
-
1
Feedback
Search any
task
Search any
task