Jailbreak Defense on Llama Model Jailbreak Evaluation Prompts
3ASRPatcher
Evaluation Results
| Method | Links | |
|---|---|---|
| PatcherAttack Type=TAP2026.06 | 3 | |
| UnpatchedAttack Type=TAP2026.06 | 4 | |
| OneShotAttack Type=TAP2026.06 | 4 | |
| FTℓ∞Attack Type=TAP2026.06 | 5 | |
| MudjackingAttack Type=TAP2026.06 | 5 | |
| MENDAttack Type=TAP2026.06 | 5 | |
| PatcherAttack Type=Puzzler2026.06 | 5 | |
| FTℓ2Attack Type=TAP2026.06 | 6 | |
| ROMEAttack Type=TAP2026.06 | 6 | |
| UnpatchedAttack Type=Puzzler2026.06 | 6 | |
| OneShotAttack Type=Puzzler2026.06 | 6 | |
| PatcherAttack Type=PAIR2026.06 | 7 | |
| FTAttack Type=TAP2026.06 | 7 | |
| SPPAttack Type=TAP2026.06 | 7 | |
| MENDAttack Type=Puzzler2026.06 | 7 | |
| UnpatchedAttack Type=PAIR2026.06 | 8 | |
| FinePruningAttack Type=TAP2026.06 | 8 | |
| FTℓ∞Attack Type=Puzzler2026.06 | 8 | |
| MudjackingAttack Type=Puzzler2026.06 | 8 | |
| OneShotAttack Type=PAIR2026.06 | 9 | |
| BAERASERAttack Type=TAP2026.06 | 9 | |
| ROMEAttack Type=Puzzler2026.06 | 9 | |
| PatcherAttack Type=GCG2026.06 | 9 | |
| MENDAttack Type=PAIR2026.06 | 10 | |
| FTℓ2Attack Type=Puzzler2026.06 | 10 | |
| UnpatchedAttack Type=GCG2026.06 | 10 | |
| OneShotAttack Type=GCG2026.06 | 10 | |
| MudjackingAttack Type=PAIR2026.06 | 11 | |
| FTAttack Type=Puzzler2026.06 | 11 | |
| SPPAttack Type=Puzzler2026.06 | 11 | |
| MENDAttack Type=GCG2026.06 | 11 | |
| PatcherAttack Type=AutoDAN2026.06 | 11 | |
| FTℓ∞Attack Type=PAIR2026.06 | 12 | |
| ROMEAttack Type=PAIR2026.06 | 12 | |
| FinePruningAttack Type=Puzzler2026.06 | 12 | |
| MudjackingAttack Type=GCG2026.06 | 12 | |
| UnpatchedAttack Type=AutoDAN2026.06 | 12 | |
| OneShotAttack Type=AutoDAN2026.06 | 12 | |
| FTℓ2Attack Type=PAIR2026.06 | 13 | |
| BAERASERAttack Type=Puzzler2026.06 | 13 | |
| FTℓ∞Attack Type=GCG2026.06 | 13 | |
| MENDAttack Type=AutoDAN2026.06 | 13 | |
| FTAttack Type=PAIR2026.06 | 14 | |
| SPPAttack Type=PAIR2026.06 | 14 | |
| ROMEAttack Type=GCG2026.06 | 14 | |
| MudjackingAttack Type=AutoDAN2026.06 | 14 | |
| FinePruningAttack Type=PAIR2026.06 | 15 | |
| FTℓ2Attack Type=GCG2026.06 | 15 | |
| FTℓ∞Attack Type=AutoDAN2026.06 | 15 | |
| BAERASERAttack Type=PAIR2026.06 | 16 | |
| FTAttack Type=GCG2026.06 | 16 | |
| SPPAttack Type=GCG2026.06 | 16 | |
| ROMEAttack Type=AutoDAN2026.06 | 16 | |
| FinePruningAttack Type=GCG2026.06 | 18 | |
| FTℓ2Attack Type=AutoDAN2026.06 | 18 | |
| BAERASERAttack Type=GCG2026.06 | 19 | |
| FTAttack Type=AutoDAN2026.06 | 19 | |
| SPPAttack Type=AutoDAN2026.06 | 19 | |
| FinePruningAttack Type=AutoDAN2026.06 | 21 | |
| BAERASERAttack Type=AutoDAN2026.06 | 22 |