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

Safety dataset

Benchmarks

Task NameDataset NameSOTA ResultTrend
Safety Classification7-class safety dataset (test)
Accuracy98.2
5
Gradient Inversion ResistanceSafety dataset 7-class (train/test)
PSNR (dB)31.7
5
Behavioral Reranking461-prompts safety dataset (test)
Baseline HCR39.5
2
Showing 3 of 3 rows