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

Harmfulness Evaluation on AJailBench Harmful

49ASR

MDSteer-c2r

-0.412.42525.2538.075Oct 20, 2025
Updated 26d ago

Evaluation Results

MethodLinks
2025.10
49
2025.10
48.76
2025.10
43.5
2025.10
38.5
2025.10
24
2025.10
19
2025.10
19
2025.10
18
2025.10
17
2025.10
12.5
2025.10
11
2025.10
1.5