Toxic Comment Detection on Toxic Comment
47.3ASRCriteria Attack
Evaluation Results
| Method | Links | |
|---|---|---|
| Criteria AttackDefense=StruQ2026.01 | 47.3 | |
| CombinedDefense=StruQ2026.01 | 18.1 | |
| TopicDefense=SecAlign2026.01 | 15.4 | |
| Criteria AttackDefense=SecAlign2026.01 | 15.2 | |
| IgnoreDefense=StruQ2026.01 | 14.2 | |
| TopicDefense=StruQ2026.01 | 12.3 | |
| Escape SeparationDefense=StruQ2026.01 | 12.1 | |
| Fake CompletionDefense=StruQ2026.01 | 11.9 | |
| IgnoreDefense=SecAlign2026.01 | 9.1 | |
| Separator InjectionDefense=StruQ2026.01 | 8.8 | |
| Escape SeparationDefense=SecAlign2026.01 | 4.7 | |
| CombinedDefense=SecAlign2026.01 | 2.1 | |
| Fake CompletionDefense=SecAlign2026.01 | 1.4 | |
| Separator InjectionDefense=SecAlign2026.01 | 1.4 |