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

Safety Evaluation on 8 jailbreak attacks (Aggregated)

3.13Average ASR

REPBEND

0.829616.357331.88547.4127Apr 2, 2025
Updated 1mo ago

Evaluation Results

MethodLinks
2025.04
3.1383.14
2025.04
3.2481.23
2025.04
5.8374.52
2025.04
7.9675.87
2025.04
8.5681.61
2025.04
9.1977.34
2025.04
10.6271.65
2025.04
10.7872.9
2025.04
11.773.62
2025.04
16.1971.85
2025.04
23.0972.67
2025.04
23.1472.96
2025.04
29.6970.71
2025.04
3473.9
2025.04
60.6463.81