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

Jailbreak Attack on DeepSeek-7b five finetuned variants

3.8Average ASR

L4A

-0.04825.92651.977.874Dec 14, 2025
Updated 1mo ago

Evaluation Results

MethodLinks
2025.12
3.8
2025.12
11
2025.12
11.6
2025.12
14.4
2025.12
23
2025.12
24.4
2025.12
27.4
2025.12
45.8
2025.12
50.4
2025.12
53.8
2025.12
59.2
2025.12
67
2025.12
71.4
2025.12
86.6
2025.12
87
2025.12
100