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

Jailbreak Defense on AutoDAN

0ASR

SafeDecoding

-3.7621.624772.38Feb 14, 2024Jun 25, 2024Nov 4, 2024Mar 16, 2025Jul 26, 2025Dec 5, 2025Apr 17, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2024.02
01.08
2024.02
01.04
2024.02
01
2024.02
01.06
2024.02
01
2024.02
01
2025.05
01
2025.05
01
2025.05
01
2026.04
0-
2026.04
0.67-
2025.05
11.02
2024.02
21.08
2024.02
21.04
2025.05
21.08
2025.05
21
2026.04
2-
2026.01
3.46-
2024.02
41.14
2025.05
41.12
2026.01
5.77-
2026.01
5.96-
2025.05
61.22
2026.01
6.54-
2026.01
6.54-
2026.01
6.92-
2026.01
7.12-
2026.01
7.12-
2026.01
7.5-
2026.01
7.88-
2025.05
81.28
2026.01
8.85-
2026.01
9.23-
2024.02
101.14
2025.05
101.36
2025.05
141.48
2026.01
18.65-
2026.01
20.77-
2025.05
221.18
2025.05
241.86
2025.05
261.5
2025.05
383.62
2025.05
382.5
2026.04
45.7-
2025.05
582.22
2025.05
663.28
2024.02
703.32
2024.02
704.64
2024.02
762.62
2024.02
804.5
2025.05
824.1
2024.02
884.92
2024.02
884.92
2025.05
884.74
2025.05
944.64