Safety Classification on Human-Authored Self-harm (test)
0.07Violation RateLlamaGuard-3-8B
Evaluation results
| Method | Links | |
|---|---|---|
| LlamaGuard-3-8BParameters=8B2026.05 | 0.07 | |
| ShieldGemma-9BParameters=9B2026.05 | 0.6 | |
| Qwen3Guard-4BParameters=4B2026.05 | 2.08 |
| Method | Links | |
|---|---|---|
| LlamaGuard-3-8BParameters=8B2026.05 | 0.07 | |
| ShieldGemma-9BParameters=9B2026.05 | 0.6 | |
| Qwen3Guard-4BParameters=4B2026.05 | 2.08 |