Harmful content detection on Trolling-oriented generations GPT-4o
19.88AccuracyPerspective API
Evaluation Results
| Method | Links | |
|---|---|---|
| Perspective API2026.04 | 19.88 | |
| OpenAI Moderation2026.04 | 18.25 | |
| LlamaGuard-22026.04 | 10.2 | |
| LlamaGuard-12026.04 | 5.65 |
Loading the SOTA2 catalog…