Harmful content detection on CADD Llama-3.1 generations
69.18AccuracyPerspective API
Evaluation Results
| Method | Links | |
|---|---|---|
| Perspective API2026.04 | 69.18 | |
| OpenAI Moderation2026.04 | 60.13 | |
| LlamaGuard-12026.04 | 45.34 | |
| LlamaGuard-22026.04 | 9.09 |
Loading the SOTA2 catalog…