Jailbreak Attack
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
99.5Safe2Harm ASR (Answer)
10
Feb 26, 2026
97.98Safe2Harm ASR (Answer)
10
Feb 26, 2026
88.1ASR
10
May 20, 2026
0.58ASR
10
Feb 26, 2026
95ASR (Refusal)
10
Feb 26, 2026
32ASR (Refusal)
10
Feb 26, 2026
61ASR
9
Jun 9, 2026
69.1ASR
9
Jun 9, 2026
55ASR (K=2)
9
May 28, 2026
92.5ASR (%) (LLaMA-3-8B)
9
May 27, 2026
6Attack Success Rate (ASR)
9
Apr 20, 2026
0ASR
9
Apr 20, 2026
0ASR
9
Apr 20, 2026
3Attack Success Rate (ASR)
9
Apr 20, 2026
1Attack Success Rate (ASR)
9
Apr 20, 2026
2.32ASR
9
Mar 25, 2026
89.8ASR
9
Mar 25, 2026
1.2ASR
9
Mar 25, 2026
92ASR
9
Mar 25, 2026
10.8ASR
9
Mar 25, 2026
98ASR
9
Mar 25, 2026
53.8ASR
9
Feb 26, 2026
100EM Rate
9
Feb 26, 2026
9,000ASR
9
Feb 26, 2026
100AK
9
Feb 26, 2026