Jailbreak Attack on ASR Benchmark
97.2ASRPuzzler
Evaluation Results
| Method | Links | |
|---|---|---|
| PuzzlerTarget Model=QwQ-32B2025.09 | 97.2 | |
| PuzzlerTarget Model=Gemma-3-1b-it2025.09 | 94 | |
| PuzzlerTarget Model=Gemini-2.0-Flash-Lite2025.09 | 86.1 | |
| PuzzlerTarget Model=Average2025.09 | 85.7 | |
| NeuroStrikeTarget Model=Gemma-3-1b-it2025.09 | 79.9 | |
| NeuroStrikeTarget Model=QwQ-32B2025.09 | 78.9 | |
| PuzzlerTarget Model=Gemini-1.5-Pro2025.09 | 75 | |
| PuzzlerTarget Model=Gemini-2.0-Flash2025.09 | 73 | |
| NeuroStrikeTarget Model=Average2025.09 | 63.7 | |
| NeuroStrikeTarget Model=Gemini-1.5-Pro2025.09 | 55.7 | |
| PAIRTarget Model=Gemini-1.5-Pro2025.09 | 54.9 | |
| NeuroStrikeTarget Model=Gemini-2.0-Flash2025.09 | 54.7 | |
| NeuroStrikeTarget Model=Gemini-2.0-Flash-Lite2025.09 | 49.2 | |
| PAIRTarget Model=Gemini-2.0-Flash2025.09 | 37.3 | |
| PAIRTarget Model=Gemma-3-1b-it2025.09 | 33.9 | |
| TAPTarget Model=Gemini-1.5-Pro2025.09 | 32 | |
| PAIRTarget Model=Average2025.09 | 31 | |
| TAPTarget Model=Gemma-3-1b-it2025.09 | 20.4 | |
| PAIRTarget Model=QwQ-32B2025.09 | 18.9 | |
| TAPTarget Model=Average2025.09 | 17.3 | |
| TAPTarget Model=Gemini-2.0-Flash2025.09 | 14 | |
| TAPTarget Model=QwQ-32B2025.09 | 12.2 | |
| PAIRTarget Model=Gemini-2.0-Flash-Lite2025.09 | 9.8 | |
| TAPTarget Model=Gemini-2.0-Flash-Lite2025.09 | 8 |