Jailbreak Evaluation
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
Jailbreak Evaluation GPT-4o-mini
93.33ASR
13
Jailbreak Evaluation Average across models
88.1ASR
10
Jailbreak Evaluation DeepSeek-V3
41.7ASR
5
Jailbreak Evaluation Qwen3
94.5Attack Success Rate (ASR)
5
Jailbreak Evaluation Gemini-2-Flash
37.5ASR
5