Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Multi-turn Jailbreak on HarmBench
Loading...
100
ASR (X-Teaming)
No Defense
-2.336
24.232
50.8
77.368
Jun 1, 2026
ASR (X-Teaming)
ASR (Tempest)
Updated 1d ago
Evaluation Results
Method
Method
Links
ASR (X-Teaming)
ASR (Tempest)
No Defense
Model=Qwen2.5-7B, Defe...
2026.06
100
84.3
No Defense
Model=Llama3-8B, Defen...
2026.06
86.7
84.7
SAGE
Model=Qwen2.5-7B, Defe...
2026.06
86.2
4.9
PROACT
Model=Qwen2.5-7B, Defe...
2026.06
63.5
2.6
PROACT
Model=Llama3-8B, Defen...
2026.06
47.4
2.4
SAGE
Model=Llama3-8B, Defen...
2026.06
17.5
6.7
THRD
Model=Qwen2.5-7B, Defe...
2026.06
4
2.1
THRD
Model=Llama3-8B, Defen...
2026.06
1.6
0.6
Feedback
Search any
task
Search any
task