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

Over-refusal Evaluation on XSTest Safe

1.6Over-refusal Rate

AdaCD

-0.3212.6425.638.56Mar 9, 2026Mar 15, 2026Mar 22, 2026Mar 29, 2026Apr 4, 2026Apr 11, 2026Apr 18, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.04
1.6
2026.04
2
2026.04
2
2026.04
2.8
2026.04
3.2
2026.03
3.6
2026.04
3.6
2026.04
4
2026.04
4.8
2026.03
5.2
2026.04
5.2
2026.04
6
2026.03
8.4
2026.04
10.4
2026.04
10.4
2026.03
11.2
2026.04
11.2
2026.04
12
2026.04
14.8
2026.04
16.4
2026.04
17.6
2026.03
23.6
2026.03
28
2026.03
28
2026.04
49.6