Harmful Prompt Detection on ViSU-sp
96PrecisionDetoxify (Orig)
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Detoxify (Orig)Zero-shot=true2026.04 | 96 | 7 | 14 | |
| DiffGuardZero-shot=true2026.04 | 92 | 15 | 25 | |
| HyPEZero-shot=true2026.04 | 73 | 90 | 81 | |
| Latent GuardZero-shot=true2026.04 | 65 | 38 | 48 | |
| NSFW-ClassifierZero-shot=true2026.04 | 59 | 58 | 58 |