| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| POPE (test) | T-CAA | CAE0.2477 | 69 | 3mo ago | |
| POPE | T-CAA | CAE27.48 | 30 | 3mo ago | |
| JailBreakV-28K | NeuroStrike | ASR (with SN)78.9 | 11 | 2mo ago | |
| GPT-5 | SRA+ | Attack Success Rate18.67 | 9 | 22d ago | |
| GPT 4.1 | HJ | Attack Success Rate (ASR)4.33 | 9 | 22d ago | |
| Safety Neuron Transfer Attack Evaluation distilled vs base | NeuroStrike | ASR Before Distillation8.6 | 5 | 2mo ago |