Multi-label Safety Categorization on HarmBench prompts
0.5432Macro AccuracyOpir-multitask-large
Evaluation Results
| Method | Links | |
|---|---|---|
| Opir-multitask-largeVariant=large2026.05 | 0.5432 | |
| Opir-multitask-multilangVariant=multilang2026.05 | 0.4828 | |
| Gliner-Guard-Omni2026.05 | 0.2986 | |
| GLiGuardParameters=300M2026.05 | 0.171 |