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

Prompt Harmfulness Classification on WildG

96.2F1 Score

HaloGuard 1.0

52.62463.93775.2586.563Jul 2, 2026Jul 3, 2026Jul 4, 2026Jul 5, 2026Jul 6, 2026Jul 7, 2026
Updated 17d ago

Evaluation Results

MethodLinks
2026.07
96.2
2026.07
95.9
2026.07
89.1
2026.07
88.9
2026.07
88.9
2026.07
88.8
2026.07
88.8
2026.07
88.7
2026.07
88.4
2026.07
88.1
2026.07
87.9
2026.07
87.8
2026.07
87.7
2026.07
87.7
2026.07
87
2026.07
87
2026.07
86.4
2026.07
85.6
2026.07
82.9
2026.07
81.6
2026.07
73
2026.07
54.3