List your company

Harmful prompt detection on I2P*

66AccuracyHyPE
0.4817.4934.551.51Apr 7, 2026

Evaluation results

MethodLinks
HyPE2026.0466
NSFW-Classifier2026.0465
Latent Guard2026.0435
DiffGuard2026.0428
GuardT2I2026.0426
Detoxify (Orig)2026.043