Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Jailbreak Attack on Advbench-M

0Attack Success Rate (ASR%)

Vanilla

-3.68421.18346.0570.917Mar 8, 2026Mar 14, 2026Mar 20, 2026Mar 26, 2026Apr 1, 2026Apr 7, 2026Apr 14, 2026
Updated 3d ago

Evaluation Results

MethodLinks
2026.03
00
2026.03
00.8
2026.04
0-
2026.04
0-
2026.04
0-
2026.04
0-
2026.04
0-
2026.04
0.14-
2026.04
0.41-
2026.03
0.51.3
2026.03
0.51.3
2026.03
0.91.8
2026.03
0.91.5
2026.04
1.1-
2026.04
1.51-
2026.03
1.91.6
2026.03
1.92.4
2026.03
2.31.3
2026.03
2.83
2026.03
3.22.9
2026.03
3.72.8
2026.03
5.11.7
2026.03
5.11.9
2026.03
5.12.1
2026.03
7.42.8
2026.03
7.91.8
2026.03
8.32.5
2026.03
10.21.5
2026.03
10.71.7
2026.03
11.12.9
2026.03
15.72.4
2026.03
16.22.3
2026.04
17.56-
2026.03
18.52.9
2026.03
19.92
2026.03
21.32.3
2026.03
22.73.3
2026.03
24.52.1
2026.03
32.92.5
2026.03
33.33.1
2026.03
47.72.8
2026.03
48.64.1
2026.03
49.15.4
2026.03
49.55.8
2026.03
52.36.1
2026.03
54.23.4
2026.03
59.35.9
2026.03
62.55.8
2026.03
64.45
2026.03
65.35.5
2026.03
695.3
2026.03
695.6
2026.03
69.97
2026.03
70.46.7
2026.03
70.46.8
2026.03
70.86.6
2026.03
73.66.5
2026.03
84.76
2026.03
86.15.5
2026.03
86.66.4
2026.03
875.8
2026.03
88.46.7
2026.03
89.85.1
2026.03
92.15.4