| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Jailbreak Defense | AutoDAN | ASR0 | 51 | |
| Jailbreak Detection | AutoDAN | Accuracy99 | 30 | |
| Jailbreak Attack | AutoDAN | ASR0.86 | 27 | |
| Adversarial Robustness | AutoDAN | ASR0 | 18 | |
| Defense Robustness | AutoDAN Attack | Keyword Success Rate100 | 16 | |
| Jailbreak Detection | AutoDAN | Attack Success Rate (ASR)0 | 15 | |
| Jailbreak Defense | AutoDAN AdvE | ASR9.83 | 14 | |
| Jailbreak Defense | AutoDAN | ASR0 | 13 | |
| Explanation Faithfulness | AutoDAN (test) | Label Change Rate8 | 8 | |
| Safety Monitoring | AutoDAN | ASR @ 1% FPR67.86 | 5 | |
| Jailbreak Resistance | AutoDAN | Refusal Rate59 | 3 | |
| Jailbreak Detection | AutoDAN-style Benchmarks | ASR (Recall)98 | 3 |