Stereotyping Evaluation on RealToxicityPrompts Non-toxic
2.8Stereotype FractionGPT-4o-mini
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| GPT-4o-miniModel=GPT-4o-m2024.07 | 2.8 | 0.57 | 0.377 | |
| GPT-4oModel=GPT-4o2024.07 | 2.9 | 0.657 | 0.356 | |
| Gemini-Flash-LiteModel=Gem-Fl-Lt2024.07 | 3.2 | 0.676 | 0.367 | |
| Gemini-ProModel=Gem-Pro2024.07 | 4.8 | 0.657 | 0.305 | |
| Gemini-FlashModel=Gem-Fl2024.07 | 5 | 0.827 | 0.406 |