Multi-turn Jailbreaking Attack on MM-SafetyBench Turn 2
24.42ASRMulti-turn Jailbreaking Attack
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Multi-turn Jailbreaking AttackTarget Model=LLaVa-13B, Toxicity Judge=Mistral-7B and Gemini 2.5-Falsh2026.01 | 24.42 | 2.3 | |
| Multi-turn Jailbreaking AttackTarget Model=Qwen-7B, Toxicity Judge=Mistral-7B and Gemini 2.5-Falsh2026.01 | 25 | 1.98 | |
| Multi-turn Jailbreaking AttackTarget Model=GPT-4o, Toxicity Judge=Mistral-7B and Gemini 2.5-Falsh2026.01 | 35.19 | 2.79 | |
| Multi-turn Jailbreaking AttackTarget Model=Gemini-2.0-Flash, Toxicity Judge=Mistral-7B and Gemini 2.5-Falsh2026.01 | 43.71 | 3.04 | |
| Multi-turn Jailbreaking AttackTarget Model=LLaVa-7B, Toxicity Judge=Mistral-7B and Gemini 2.5-Falsh2026.01 | 73.08 | 3.75 |