Our new X account is live! Follow @wizwand_team for updates
Home
/
Benchmarks
Jailbreak Attack Success Rate on HarmBench (50 randomly sampled questions)
Loading...
84
ASR
GPT-OSS-20B
-3.36
19.32
42
64.68
Feb 3, 2026
ASR
Updated 4d ago
Evaluation Results
Method
Method
Links
ASR
GPT-OSS-20B
Steering Condition=STE...
2026.02
84
Llama-3-8B-Instruct-RR
Steering Condition=STE...
2026.02
70
GPT-OSS-20B
Steering Condition=Non...
2026.02
62
Llama-3-8B-Instruct-RR
Steering Condition=Non...
2026.02
52
GPT-OSS-20B
Steering Condition=STE...
2026.02
6
Llama-3-8B-Instruct-RR
Steering Condition=STE...
2026.02
2
Llama-3-8B-Instruct-RR
Steering Condition=Non...
2026.02
0
GPT-OSS-20B
Steering Condition=Non...
2026.02
0
Feedback
Search any
task
Search any
task