Content Safety Detection on Aegis-Safety 2.0
95.1F1 Scoreprompt–response verification framework
Evaluation results
| Method | Links | ||
|---|---|---|---|
| prompt–response verification framework2026.06 | 95.1 | 4.1 | |
| WildGuard / ShieldGemmaType=Strongest Applicable Baseline2026.06 | 89.5 | — |