Content Safety on ChineseHarm (test)
97.9Macro F1-ScoreYuvion-32B
Evaluation Results
| Method | Links | |
|---|---|---|
| Yuvion-32BParameters=32B2026.06 | 97.9 | |
| Yuvion-8BParameters=8B2026.06 | 96.8 | |
| Qwen2-32BParameters=32B2026.06 | 93.2 | |
| Qwen2-8BParameters=8B2026.06 | 90.8 | |
| Qwen2-72BParameters=72B2026.06 | 87.9 | |
| GPT-3.52026.06 | 85.3 | |
| Qwen1.5-72BParameters=72B2026.06 | 81.2 | |
| DeepSeek-67BParameters=67B2026.06 | 54.4 |