| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| PKU-SafeRLHF (test) | SafeMoE-XL | Drugs & Weapons Safety Score85.3 | 6 | 1d ago | |
| OrBench Hard | Mistral MoE-XL | Deception Rate (Safe)75 | 4 | 1d ago | |
| HarmfulQA Social science | Mistral MoE-XL | Safety Score90 | 4 | 1d ago | |
| HarmfulQA Science and Technology | Mistral MoE-XL | Safety Score95 | 4 | 1d ago | |
| HarmfulQA Philosophy and Ethics | Mistral MoE-XL | Safety Score80 | 4 | 1d ago | |
| HarmfulQA Mathematics and Logic | Qwen MoE-XL | Safety Score76.7 | 4 | 1d ago | |
| HarmfulQA Literature and Language | Mistral MoE-XL | Safety Score100 | 4 | 1d ago | |
| HarmfulQA History and Culture | Mistral MoE-XL | Safety Score90 | 4 | 1d ago | |
| HarmfulQA Health and Medicine | Mistral MoE-XL | Safety Score85 | 4 | 1d ago | |
| HarmfulQA Geography and Environment | Mistral MoE-XL | Safety Rate95 | 4 | 1d ago | |
| HarmfulQA Education and Pedagogy | Mistral MoE-XL | Safety Score100 | 4 | 1d ago | |
| HarmfulQA Business and Economic | Mistral MoE-XL | Safety Rate91 | 4 | 1d ago |