Model Benchmarks
Compare model performance across safety dimensions.
Performance by category
PolicyViolenceSelf-harmSexualPrivacy
| Category | Accuracy | Recall | Precision |
|---|---|---|---|
| Violence | 98.1% | 97.0% | 96.4% |
| Self-harm | 97.6% | 96.1% | 95.9% |
| Privacy | 98.4% | 97.2% | 96.7% |
Compare model performance across safety dimensions.
| Category | Accuracy | Recall | Precision |
|---|---|---|---|
| Violence | 98.1% | 97.0% | 96.4% |
| Self-harm | 97.6% | 96.1% | 95.9% |
| Privacy | 98.4% | 97.2% | 96.7% |