Jailbreak Detection on Zulu
92AccuracyIMAG
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| IMAGTarget LLM=Mistral-7B2025.12 | 92 | 96 | |
| IMAGTarget LLM=Vicuna-13B2025.12 | 65 | 79 | |
| LlamaGTarget LLM=Mistral-7B2025.12 | 58 | 73 | |
| LlamaGTarget LLM=Vicuna-7B2025.12 | 55 | 71 | |
| LlamaGTarget LLM=Vicuna-13B2025.12 | 54 | 70 | |
| LlamaGTarget LLM=Llama3-8B2025.12 | 38 | 56 | |
| Self-ExTarget LLM=Mistral-7B2025.12 | 37 | 54 | |
| LlamaGTarget LLM=Llama2-7B2025.12 | 35 | 51 | |
| IMAGTarget LLM=Llama3-8B2025.12 | 32 | 48 | |
| Self-ExTarget LLM=Llama2-7B2025.12 | 20 | 32 | |
| IMAGTarget LLM=Vicuna-7B2025.12 | 19 | 32 | |
| IMAGTarget LLM=Llama2-7B2025.12 | 19 | 32 | |
| GradSafeTarget LLM=Llama2-7B2025.12 | 18 | 31 | |
| Self-ExTarget LLM=Llama3-8B2025.12 | 14 | 24 | |
| Self-ExTarget LLM=Vicuna-7B2025.12 | 1 | 3 | |
| OAPITarget LLM=Mistral-7B2025.12 | 0 | 1 | |
| PPLTarget LLM=Mistral-7B2025.12 | 0 | 0 | |
| GradSafeTarget LLM=Mistral-7B2025.12 | 0 | 0 | |
| OAPITarget LLM=Vicuna-7B2025.12 | 0 | 0 | |
| PPLTarget LLM=Vicuna-7B2025.12 | 0 | 0 | |
| GradSafeTarget LLM=Vicuna-7B2025.12 | 0 | 0 | |
| OAPITarget LLM=Vicuna-13B2025.12 | 0 | 0 | |
| PPLTarget LLM=Vicuna-13B2025.12 | 0 | 0 | |
| Self-ExTarget LLM=Vicuna-13B2025.12 | 0 | 0 | |
| GradSafeTarget LLM=Vicuna-13B2025.12 | 0 | 0 | |
| OAPITarget LLM=Llama2-7B2025.12 | 0 | 0 | |
| PPLTarget LLM=Llama2-7B2025.12 | 0 | 0 | |
| OAPITarget LLM=Llama3-8B2025.12 | 0 | 0 | |
| PPLTarget LLM=Llama3-8B2025.12 | 0 | 0 | |
| GradSafeTarget LLM=Llama3-8B2025.12 | 0 | 0 |