Multi-label Safety Categorization on HarmBench Responses
27.26Macro AccuracyOpir-multitask-large
Evaluation Results
| Method | Links | |
|---|---|---|
| Opir-multitask-largeVariant=large2026.05 | 27.26 | |
| Opir-multitask-multilangVariant=multilang2026.05 | 21.58 | |
| GLiGuardParameters=300M2026.05 | 20.09 | |
| Gliner-Guard-Omni2026.05 | 2.94 |