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

Jailbreak Attack Success Rate on SorryBench

8.9ASR (SorryBench)

ChatBug-M

5.28429.69254.178.508Mar 31, 2025May 2, 2025Jun 3, 2025Jul 6, 2025Aug 7, 2025Sep 8, 2025Oct 11, 2025
Updated 26d ago

Evaluation Results

MethodLinks
2025.10
8.9-
2025.10
10.2-
2025.10
10.2-
2025.03
15.5-
2025.03
16.6-
2025.03
16.8-
2025.10
19.3-
2025.10
24.3-
2025.03
29.3-
2025.03
29.3-
2025.10
29.8-
2025.10
30.2-
2025.10
31.1-
2025.10
31.6-
2025.10
32.3-
2025.10
32.3-
2025.03
33.9-
2025.03
33.9-
2025.10
33.9-
2025.10
37-
2025.10
37.7-
2025.10
37.7-
2025.10
37.7-
2025.10
42-
2025.03
42.5-
2025.03
42.5-
2025.10
45.2-
2025.10
45.5-
2025.10
53-
2025.10
58.6-
2025.10
60.9-
2025.10
62.5-
2025.10
67.7-
2025.10
69.127.1
2025.10
69.135.2
2025.10
69.8-
2025.10
71.1-
2025.10
81.814.1
2025.10
82.3-
2025.10
83.412.3
2025.10
83.420.9
2025.03
83.8-
2025.10
8524.1
2025.03
85.2-
2025.10
86.833.8
2025.10
87.3-
2025.03
89.5-
2025.03
91.6-
2025.10
92.710.4
2025.03
93.2-
2025.03
93.9-
2025.03
94.8-
2025.03
95.5-
2025.03
96.4-
2025.03
96.4-
2025.03
96.4-
2025.03
96.9-
2025.03
96.9-
2025.03
97.6-
2025.03
98.6-
2025.03
98.6-
2025.03
99.3-