Indirect Prompt Injection Defense on Ignore Instruction
0ASRLLM-Detector
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| LLM-DetectorFoundation Model=QWEN-3-8B, Base Utility=69.82026.02 | 0 | 47.2 | |
| ICONFoundation Model=QWEN-3-8B, Base Utility=69.82026.02 | 0 | 58.2 | |
| LLM-DetectorFoundation Model=LLAMA-3.1-8B, Base Utility=34.52026.02 | 0 | 25.2 | |
| ICONFoundation Model=LLAMA-3.1-8B, Base Utility=34.52026.02 | 0 | 32.6 | |
| LLM-DetectorFoundation Model=MISTRAL-8B, Base Utility=20.12026.02 | 0 | 11.4 | |
| ICONFoundation Model=MISTRAL-8B, Base Utility=20.12026.02 | 0 | 17.6 | |
| MELONFoundation Model=MISTRAL-8B, Base Utility=20.12026.02 | 0.2 | 12.1 | |
| MELONFoundation Model=LLAMA-3.1-8B, Base Utility=34.52026.02 | 0.6 | 26.1 | |
| MELONFoundation Model=QWEN-3-8B, Base Utility=69.82026.02 | 1.3 | 48.7 | |
| Repeat PromptFoundation Model=MISTRAL-8B, Base Utility=20.12026.02 | 10.3 | 17.8 | |
| Repeat PromptFoundation Model=QWEN-3-8B, Base Utility=69.82026.02 | 11.7 | 54.2 | |
| DelimitingFoundation Model=MISTRAL-8B, Base Utility=20.12026.02 | 12.1 | 16.5 | |
| Repeat PromptFoundation Model=LLAMA-3.1-8B, Base Utility=34.52026.02 | 14.6 | 30.1 | |
| DelimitingFoundation Model=QWEN-3-8B, Base Utility=69.82026.02 | 15.3 | 52.9 | |
| No DefenseFoundation Model=MISTRAL-8B, Base Utility=20.12026.02 | 16.3 | 12.5 | |
| DelimitingFoundation Model=LLAMA-3.1-8B, Base Utility=34.52026.02 | 16.8 | 28.1 | |
| No DefenseFoundation Model=LLAMA-3.1-8B, Base Utility=34.52026.02 | 23.4 | 25.7 | |
| No DefenseFoundation Model=QWEN-3-8B, Base Utility=69.82026.02 | 25 | 49.7 |