Jailbreak Defense on AdvBench DIJA
6.94ASRDream-v0-Instruct-7B + DIFFUGUARD
Evaluation Results
| Method | Links | |
|---|---|---|
| Dream-v0-Instruct-7B + DIFFUGUARDBase Model=Dream-v0-Instruct-7B, Defense Method=DIFFUGUARD2025.09 | 6.94 | |
| Dream-v0-Instruct-7B + Self-reminder + DIFFUGUARDBase Model=Dream-v0-Instruct-7B, Defense Method=Self-reminder + DIFFUGUARD2025.09 | 14.26 | |
| LLaDA-1.5 + Self-reminder + DIFFUGUARDBase Model=LLaDA-1.5, Defense Method=Self-reminder + DIFFUGUARD2025.09 | 19.04 | |
| LLaDA-8B-Instruct + Self-reminder + DIFFUGUARDBase Model=LLaDA-8B-Instruct, Defense Method=Self-reminder + DIFFUGUARD2025.09 | 39.04 | |
| MMaDA-8B-MixCoT + Self-reminder + DIFFUGUARDBase Model=MMaDA-8B-MixCoT, Defense Method=Self-reminder + DIFFUGUARD2025.09 | 39.23 | |
| LLaDA-1.5 + DIFFUGUARDBase Model=LLaDA-1.5, Defense Method=DIFFUGUARD2025.09 | 51.54 | |
| LLaDA-8B-Instruct + DIFFUGUARDBase Model=LLaDA-8B-Instruct, Defense Method=DIFFUGUARD2025.09 | 51.92 | |
| MMaDA-8B-MixCoT + DIFFUGUARDBase Model=MMaDA-8B-MixCoT, Defense Method=DIFFUGUARD2025.09 | 53.85 | |
| LLaDA-1.5 + PPL-FilterBase Model=LLaDA-1.5, Defense Method=PPL-Filter2025.09 | 83.85 | |
| MMaDA-8B-MixCoT + PPL-FilterBase Model=MMaDA-8B-MixCoT, Defense Method=PPL-Filter2025.09 | 89.81 | |
| LLaDA-8B-Instruct + PPL-FilterBase Model=LLaDA-8B-Instruct, Defense Method=PPL-Filter2025.09 | 90.19 | |
| Dream-v0-Instruct-7B + PPL-FilterBase Model=Dream-v0-Instruct-7B, Defense Method=PPL-Filter2025.09 | 91.92 | |
| LLaDA-1.5 + Self-reminderBase Model=LLaDA-1.5, Defense Method=Self-reminder2025.09 | 97.31 | |
| LLaDA-8B-Instruct + Self-reminderBase Model=LLaDA-8B-Instruct, Defense Method=Self-reminder2025.09 | 97.5 | |
| Dream-v0-Instruct-7B + Self-reminderBase Model=Dream-v0-Instruct-7B, Defense Method=Self-reminder2025.09 | 97.69 | |
| LLaDA-1.5Base Model=LLaDA-1.5, Defense Method=None2025.09 | 97.88 | |
| MMaDA-8B-MixCoT + Self-reminderBase Model=MMaDA-8B-MixCoT, Defense Method=Self-reminder2025.09 | 98.27 | |
| LLaDA-8B-InstructBase Model=LLaDA-8B-Instruct, Defense Method=None2025.09 | 98.65 | |
| MMaDA-8B-MixCoTBase Model=MMaDA-8B-MixCoT, Defense Method=None2025.09 | 98.85 | |
| Dream-v0-Instruct-7BBase Model=Dream-v0-Instruct-7B, Defense Method=None2025.09 | 99.23 |