Jailbreak Attack on MaliciousInstruct 41 (test)
0.935ASRShallowJail
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| ShallowJailVictim Model=Llama-3.1-8B-Instruct, tau=150, alpha=0.8, beta=0.52026.02 | 0.935 | 19.1552 | |
| ShallowJailVictim Model=Qwen3-4B-Instruct-2507, tau=150, alpha=5, beta=0.52026.02 | 0.795 | 4.3775 | |
| ShallowJailVictim Model=Qwen2.5-7B-Instruct, tau=150, alpha=6.5, beta=0.52026.02 | 0.765 | 5.5675 | |
| DirectVictim Model=Qwen2.5-7B-Instruct, tau=150, alpha=6.5, beta=0.52026.02 | 0.04 | 1.9173 | |
| DirectVictim Model=Llama-3.1-8B-Instruct, tau=150, alpha=0.8, beta=0.52026.02 | 0.02 | 1.5375 | |
| DirectVictim Model=Qwen3-4B-Instruct-2507, tau=150, alpha=5, beta=0.52026.02 | 0.005 | 1.506 |