Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Red-teaming on HarmBench Claude-4-Sonnet (test)
Loading...
71
ASR@1
DIALTREE
-1.28
17.485
36.25
55.015
Oct 2, 2025
ASR@1
ASR@5
Updated 1mo ago
Evaluation Results
Method
Method
Links
ASR@1
ASR@5
DIALTREE
Target Model=Claude-4-...
2025.10
71
96.5
X-Teaming
Target Model=Claude-4-...
2025.10
9.5
32.5
SFT
Target Model=Claude-4-...
2025.10
1.5
4
Feedback
Search any
task
Search any
task