Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Attack Detection on C4 Task-aligned filler (N = 100)
Loading...
10
TPR
Output-Length Anomaly
-0.4
2.3
5
7.7
Jun 6, 2026
TPR
Updated 1mo ago
Evaluation Results
Method
Method
Links
TPR
Output-Length Anomaly
Backbone Model=DS-R1-Q...
2026.06
10
Token Budget
Backbone Model=DS-R1-Q...
2026.06
3
Compression-Ratio Proxy
Backbone Model=DS-R1-Q...
2026.06
1
Answer-Absence Proxy
Backbone Model=DS-R1-Q...
2026.06
1
QDM
Backbone Model=DS-R1-Q...
2026.06
0
RECURGUARD
Backbone Model=DS-R1-Q...
2026.06
0
Feedback
Search any
task
Search any
task