Jailbreak Attack on GLM-4-Air
94NRSWhisper
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| SWhisperSurrogate Model=Gemma-3-4B2026.03 | 94 | 92.5 | |
| SWhisperSurrogate Model=LLaMA-3.1-8B2026.03 | 92 | 88.5 | |
| SWhisperSurrogate Model=Qwen2.5-7B2026.03 | 90 | 86 | |
| SWhisperSurrogate Model=Mistral-7B2026.03 | 86 | 85.3 | |
| Auto-DANSurrogate Model=Gemma-3-4B2026.03 | 68 | 68 | |
| Auto-DANSurrogate Model=Mistral-7B2026.03 | 66 | 66 | |
| Auto-DANSurrogate Model=Qwen2.5-7B2026.03 | 64 | 62.3 | |
| Auto-DANSurrogate Model=LLaMA-3.1-8B2026.03 | 56 | 55.5 | |
| PAIRSurrogate Model=Qwen2.5-7B2026.03 | 22 | 18.8 | |
| PAIRSurrogate Model=LLaMA-3.1-8B2026.03 | 18 | 15.5 | |
| PAIRSurrogate Model=Mistral-7B2026.03 | 18 | 18 | |
| PAIRSurrogate Model=Gemma-3-4B2026.03 | 14 | 13.8 | |
| IGCGSurrogate Model=LLaMA-3.1-8B2026.03 | 6 | 5.3 | |
| GCGSurrogate Model=LLaMA-3.1-8B2026.03 | 4 | 3.5 | |
| GCGSurrogate Model=Qwen2.5-7B2026.03 | 2 | 2 | |
| GCGSurrogate Model=Mistral-7B2026.03 | 2 | 2 | |
| IGCGSurrogate Model=Gemma-3-4B2026.03 | 2 | 2 | |
| IGCGSurrogate Model=Qwen2.5-7B2026.03 | 2 | 2 | |
| GCGSurrogate Model=Gemma-3-4B2026.03 | 0 | 0 | |
| IGCGSurrogate Model=Mistral-7B2026.03 | 0 | 0 |