Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

Safety Evaluation on WildJailbreak (held-out)

0Attack Success Rate (ASR)

NeST

-3.16418.19339.5560.907Feb 18, 2026
Updated 4d ago

Evaluation Results

MethodLinks
2026.02
0
2026.02
0.1
2026.02
0.3
2026.02
0.4
2026.02
0.6
2026.02
0.7
2026.02
0.7
2026.02
0.8
2026.02
1.4
2026.02
1.5
2026.02
1.6
2026.02
1.6
2026.02
1.7
2026.02
1.9
2026.02
1.9
2026.02
2.1
2026.02
2.8
2026.02
2.8
2026.02
2.9
2026.02
3.4
2026.02
4.1
2026.02
4.2
2026.02
5.7
2026.02
6.6
2026.02
6.7
2026.02
11
2026.02
12.2
2026.02
13
2026.02
17.8
2026.02
19.2
2026.02
22
2026.02
25.2
2026.02
25.7
2026.02
34
2026.02
39.7
2026.02
40.5
2026.02
40.6
2026.02
44.5
2026.02
50.9
2026.02
54.2
2026.02
58.1
2026.02
58.1
2026.02
65.1
2026.02
65.2
2026.02
68.2
2026.02
69.4
2026.02
72.1
2026.02
74.8
2026.02
76.4
2026.02
79.1