Adversarial Detection on ToxicPrompts (test)
11Inference Latency (ms)Rule-Based
Evaluation results
| Method | Links | |
|---|---|---|
| Rule-Based2026.05 | 11 | |
| APD2026.05 | 12.5 | |
| ECFull name=Embedding Clustering2026.05 | 16 | |
| ATFull name=Adversarial Training2026.05 | 38.9 | |
| Post-Output2026.05 | 46.2 |