| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| Dataset of 200 scientific papers 1.0 (test) | C1DRA86.26 | 13 | 3mo ago | ||
| POLITEHARMBENCH 1.0 (full) | ASR0 | 11 | 3mo ago | ||
| HarmBench | SafeMoE-XL | Attack Success Rate (ASR)1 | 4 | 1d ago | |
| PKU-Safe | SafeMoE-XL | Attack Success Rate (ASR)25 | 4 | 1d ago | |
| PGDJ64 | CARE | ASR8.46 | 3 | 2mo ago | |
| PGDT64 | CARE | Attack Success Rate (ASR)4.6 | 3 | 2mo ago | |
| Tiny-ImageNet | RSLAD | Clean Avg Acc47.95 | 3 | 3mo ago |