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

ProGuard

Benchmarks

Task NameDataset NameSOTA ResultTrend
Unsafe content categorizationProGuard Text
Accuracy76.96
9
Unsafe content categorizationProGuard Text-Image
Accuracy0.6997
6
Unsafe content categorizationProGuard Image
Accuracy76.02
5
OOD safety category inference (Stage 2)ProGuard Text-Image
Mean Reward26.86
4
Out-of-Taxonomy Risk DetectionProGuard Image
F1 Score57.59
4
Out-of-Taxonomy Risk DetectionProGuard Text-Image
F1 Score (%)60.25
4
Out-of-Taxonomy Risk DetectionProGuard Text
F1 Score56.94
4
OOD safety category inference (Stage 2)ProGuard Image
Mean Reward25.95
4
Showing 8 of 8 rows