Jailbreak Attack on JailbreakBench GCG
0.58ASRNone
Evaluation results
| Method | Links | |
|---|---|---|
| NoneModel=Vicuna2024.01 | 0.58 | |
| Perplexity FilterModel=Vicuna2024.01 | 0.14 | |
| NoneModel=Llama-22024.01 | 0.02 | |
| SmoothLLMModel=Vicuna2024.01 | 0.01 | |
| SmoothLLMModel=Llama-22024.01 | 0.01 | |
| Perplexity FilterModel=Llama-22024.01 | 0.01 | |
| RephrasingModel=Vicuna2024.01 | 0.01 | |
| RPOModel=Vicuna2024.01 | 0.01 | |
| RephrasingModel=Llama-22024.01 | 0 | |
| RPOModel=Llama-22024.01 | 0 |