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

Agent Safety Evaluation on AgentHarm Harmful Requests

59Score

GPT-4o-mini

-2.3613.5729.545.43Jun 1, 2026
Updated 1d ago

Evaluation Results

MethodLinks
2026.06
59286
2026.06
52286
2026.06
51345
2026.06
4796
2026.06
45336
2026.06
42.875
2026.06
3999
2026.06
36269
2026.06
362410
2026.06
1959
2026.06
1808
2026.06
1759
2026.06
13314
2026.06
12913
2026.06
10614
2026.06
8516
2026.06
8216
2026.06
7516
2026.06
6217
2026.06
2118
2026.06
2218
2026.06
2118
2026.06
2018
2026.06
1018
2026.06
0018
2026.06
0018
2026.06
0018