Jailbreak Attack on LLaMA-8B-J unseen instances
0.756HypervolumeEvoJail
Evaluation results
| Method | Links | |
|---|---|---|
| EvoJailEvaluation Level=ensemble level, Instance Type=unseen instances2026.03 | 0.756 | |
| CodeAttackEvaluation Level=ensemble level, Instance Type=unseen instances2026.03 | 0.715 | |
| JailbrokenEvaluation Level=ensemble level, Instance Type=unseen instances2026.03 | 0.584 | |
| CodeChameleonEvaluation Level=ensemble level, Instance Type=unseen instances2026.03 | 0.511 | |
| ReNeLLMEvaluation Level=ensemble level, Instance Type=unseen instances2026.03 | 0.454 | |
| FlipAttackEvaluation Level=ensemble level, Instance Type=unseen instances2026.03 | 0.445 | |
| CipherEvaluation Level=ensemble level, Instance Type=unseen instances2026.03 | 0 |