Toxicity and Harmful Content Detection on AgentHarm
94.69ScoreJT-Safe-V2-35B
Evaluation Results
| Method | Links | |
|---|---|---|
| JT-Safe-V2-35BNumber of Parameters=35B2026.05 | 94.69 | |
| Qwen3-235BNumber of Parameters=235B2026.05 | 94.62 | |
| Qwen3-32BNumber of Parameters=32B2026.05 | 92.77 | |
| DeepSeek-V3.22026.05 | 90 | |
| Qwen3.5-35BNumber of Parameters=35B2026.05 | 84.23 |