Question Answering on Bio (test)
82.9LLM-Judge ScoreInstructRAG
Evaluation Results
| Method | Links | |
|---|---|---|
| InstructRAGModel=GPT-4o-mini, Poison Position=25, k=502025.09 | 82.9 | |
| InstructRAGModel=GPT-4o-mini, Poison Position=50, k=502025.09 | 82.7 | |
| Sampling + MISModel=GPT-4o-mini, Attack Position=Benign, k=502025.09 | 81.3 | |
| Vanilla RAGModel=GPT-4o-mini, Attack Position=Benign, k=502025.09 | 81.2 | |
| AstuteRAGModel=GPT-4o-mini, Poison Position=25, k=502025.09 | 80.8 | |
| AstuteRAGModel=GPT-4o-mini, Poison Position=50, k=502025.09 | 80.7 | |
| Sampling + MISModel=GPT-4o-mini, Poison Position=50, k=502025.09 | 80.7 | |
| Vanilla RAGModel=GPT-4o-mini, Poison Position=1, k=502025.09 | 80.3 | |
| Sampling + MISModel=GPT-4o-mini, Poison Position=25, k=502025.09 | 80.2 | |
| Sampling + MISModel=GPT-4o-mini, Attack Position=@Pos 50, k=502025.09 | 78.2 | |
| AstuteRAGModel=GPT-4o-mini, Attack Position=@Pos 25, k=502025.09 | 78.1 | |
| AstuteRAGModel=GPT-4o-mini, Attack Position=@Pos 50, k=502025.09 | 77.1 | |
| Sampling + MISModel=GPT-4o-mini, Attack Position=@Pos 25, k=502025.09 | 76.7 | |
| InstructRAGModel=GPT-4o-mini, Poison Position=1, k=502025.09 | 76.1 | |
| Sampling + MISModel=GPT-4o-mini, Poison Position=1, k=502025.09 | 76.1 | |
| Sampling + MISModel=GPT-4o-mini, Attack Position=@Pos 1, k=502025.09 | 75.6 | |
| Vanilla RAGModel=GPT-4o-mini, Attack Position=@Pos 1, k=502025.09 | 75.3 | |
| InstructRAGModel=GPT-4o-mini, Attack Position=Benign, k=502025.09 | 74.7 | |
| InstructRAGModel=GPT-4o-mini, Attack Position=@Pos 25, k=502025.09 | 74.3 | |
| Vanilla RAGModel=GPT-4o-mini, Poison Position=25, k=502025.09 | 73.4 | |
| AstuteRAGModel=GPT-4o-mini, Attack Position=Benign, k=502025.09 | 73.3 | |
| Sampling + MISModel=Llama3.2-3B, Attack Position=Benign, k=502025.09 | 72 | |
| AstuteRAGModel=GPT-4o-mini, Poison Position=1, k=502025.09 | 71.8 | |
| Sampling + MISModel=Mistral-7B, Attack Position=Benign, k=502025.09 | 70.9 | |
| InstructRAGModel=Llama3.2-3B, Attack Position=@Pos 1, k=502025.09 | 69.8 | |
| Sampling + MISModel=Mistral-7B, Attack Position=@Pos 25, k=502025.09 | 69.7 | |
| Vanilla RAGModel=Mistral-7B, Poison Position=1, k=502025.09 | 69.6 | |
| Sampling + MISModel=Mistral-7B, Poison Position=25, k=502025.09 | 69.6 | |
| Sampling + MISModel=Llama3.2-3B, Attack Position=@Pos 50, k=502025.09 | 69.5 | |
| InstructRAGModel=Llama3.2-3B, Attack Position=Benign, k=502025.09 | 69.3 | |
| Vanilla RAGModel=Mistral-7B, Attack Position=Benign, k=502025.09 | 69 | |
| Sampling + MISModel=Mistral-7B, Poison Position=50, k=502025.09 | 69 | |
| Vanilla RAGModel=Llama3.2-3B, Attack Position=@Pos 1, k=502025.09 | 68.9 | |
| Vanilla RAGModel=Llama3.2-3B, Poison Position=1, k=502025.09 | 68.9 | |
| Sampling + MISModel=Llama3.2-3B, Poison Position=50, k=502025.09 | 68.9 | |
| InstructRAGModel=Mistral-7B, Attack Position=Benign, k=502025.09 | 68.7 | |
| Vanilla RAGModel=Mistral-7B, Attack Position=@Pos 1, k=502025.09 | 67.9 | |
| Vanilla RAGModel=Llama3.2-3B, Attack Position=Benign, k=502025.09 | 67.9 | |
| RobustRAGModel=GPT-4o-mini, Poison Position=50, k=502025.09 | 67.6 | |
| RobustRAGModel=GPT-4o-mini, Poison Position=25, k=502025.09 | 67.5 | |
| RobustRAGModel=GPT-4o-mini, Attack Position=@Pos 25, k=502025.09 | 67.2 | |
| RobustRAGModel=GPT-4o-mini, Poison Position=1, k=502025.09 | 67 | |
| Sampling + MISModel=Llama3.2-3B, Attack Position=@Pos 25, k=502025.09 | 66.6 | |
| RobustRAGModel=GPT-4o-mini, Attack Position=@Pos 1, k=502025.09 | 66.5 | |
| Vanilla RAGModel=GPT-4o-mini, Poison Position=50, k=502025.09 | 66.1 | |
| Sampling + MISModel=Mistral-7B, Poison Position=1, k=502025.09 | 65.9 | |
| Sampling + MISModel=Llama3.2-3B, Poison Position=25, k=502025.09 | 65.9 | |
| InstructRAGModel=GPT-4o-mini, Attack Position=@Pos 50, k=502025.09 | 65.8 | |
| RobustRAGModel=GPT-4o-mini, Attack Position=Benign, k=502025.09 | 65.7 | |
| RobustRAGModel=GPT-4o-mini, Attack Position=@Pos 50, k=502025.09 | 65.5 | |
| Sampling + MISModel=Llama3.2-3B, Poison Position=1, k=502025.09 | 65.3 | |
| InstructRAGModel=Mistral-7B, Attack Position=@Pos 1, k=502025.09 | 65.1 | |
| Sampling + MISModel=Mistral-7B, Attack Position=@Pos 50, k=502025.09 | 64.2 | |
| Sampling + MISModel=Llama3.2-3B, Attack Position=@Pos 1, k=502025.09 | 64.1 | |
| InstructRAGModel=Mistral-7B, Poison Position=1, k=502025.09 | 63.8 | |
| AstuteRAGModel=GPT-4o-mini, Attack Position=@Pos 1, k=502025.09 | 63.2 | |
| RobustRAGModel=Mistral-7B, Attack Position=Benign, k=502025.09 | 61.7 | |
| InstructRAGModel=GPT-4o-mini, Attack Position=@Pos 1, k=502025.09 | 61.6 | |
| RobustRAGModel=Mistral-7B, Poison Position=25, k=502025.09 | 61.3 | |
| AstuteRAGModel=Llama3.2-3B, Attack Position=Benign, k=502025.09 | 61 | |
| RobustRAGModel=Mistral-7B, Attack Position=@Pos 50, k=502025.09 | 60.4 | |
| InstructRAGModel=Mistral-7B, Poison Position=25, k=502025.09 | 60.4 | |
| RobustRAGModel=Mistral-7B, Poison Position=1, k=502025.09 | 60.2 | |
| RobustRAGModel=Mistral-7B, Attack Position=@Pos 25, k=502025.09 | 59.8 | |
| RobustRAGModel=Mistral-7B, Poison Position=50, k=502025.09 | 59.7 | |
| RobustRAGModel=Mistral-7B, Attack Position=@Pos 1, k=502025.09 | 59.4 | |
| InstructRAGModel=Llama3.2-3B, Poison Position=1, k=502025.09 | 59.4 | |
| AstuteRAGModel=Mistral-7B, Attack Position=Benign, k=502025.09 | 59.2 | |
| RobustRAGModel=Llama3.2-3B, Poison Position=50, k=502025.09 | 58.3 | |
| RobustRAGModel=Llama3.2-3B, Poison Position=25, k=502025.09 | 58 | |
| Sampling + MISModel=Mistral-7B, Attack Position=@Pos 1, k=502025.09 | 57.5 | |
| RobustRAGModel=Llama3.2-3B, Attack Position=@Pos 25, k=502025.09 | 57.5 | |
| RobustRAGModel=Llama3.2-3B, Attack Position=@Pos 50, k=502025.09 | 57.3 | |
| InstructRAGModel=Mistral-7B, Poison Position=50, k=502025.09 | 56.9 | |
| RobustRAGModel=Llama3.2-3B, Attack Position=@Pos 1, k=502025.09 | 56.8 | |
| AstuteRAGModel=Llama3.2-3B, Attack Position=@Pos 1, k=502025.09 | 56.6 | |
| RobustRAGModel=Llama3.2-3B, Attack Position=Benign, k=502025.09 | 56.1 | |
| RobustRAGModel=Llama3.2-3B, Poison Position=1, k=502025.09 | 56.1 | |
| AstuteRAGModel=Mistral-7B, Attack Position=@Pos 1, k=502025.09 | 55.7 | |
| Vanilla RAGModel=Mistral-7B, Poison Position=25, k=502025.09 | 53.5 | |
| AstuteRAGModel=Mistral-7B, Poison Position=1, k=502025.09 | 52.6 | |
| Vanilla RAGModel=Llama3.2-3B, Poison Position=25, k=502025.09 | 52.5 | |
| InstructRAGModel=Llama3.2-3B, Attack Position=@Pos 25, k=502025.09 | 51.9 | |
| AstuteRAGModel=Mistral-7B, Attack Position=@Pos 25, k=502025.09 | 51.3 | |
| InstructRAGModel=Llama3.2-3B, Poison Position=25, k=502025.09 | 48.2 | |
| AstuteRAGModel=Llama3.2-3B, Attack Position=@Pos 25, k=502025.09 | 47.3 | |
| AstuteRAGModel=Mistral-7B, Poison Position=25, k=502025.09 | 46.1 | |
| Vanilla RAGModel=Mistral-7B, Poison Position=50, k=502025.09 | 45.5 | |
| AstuteRAGModel=Mistral-7B, Attack Position=@Pos 50, k=502025.09 | 44.8 | |
| AstuteRAGModel=Llama3.2-3B, Poison Position=25, k=502025.09 | 44.7 | |
| AstuteRAGModel=Llama3.2-3B, Poison Position=1, k=502025.09 | 44.6 | |
| AstuteRAGModel=Mistral-7B, Poison Position=50, k=502025.09 | 44.2 | |
| Vanilla RAGModel=GPT-4o-mini, Attack Position=@Pos 25, k=502025.09 | 43.1 | |
| Vanilla RAGModel=Mistral-7B, Attack Position=@Pos 25, k=502025.09 | 40.7 | |
| AstuteRAGModel=Llama3.2-3B, Poison Position=50, k=502025.09 | 40.7 | |
| Vanilla RAGModel=Llama3.2-3B, Attack Position=@Pos 25, k=502025.09 | 40.4 | |
| InstructRAGModel=Mistral-7B, Attack Position=@Pos 25, k=502025.09 | 38.4 | |
| Vanilla RAGModel=Llama3.2-3B, Poison Position=50, k=502025.09 | 37.9 | |
| InstructRAGModel=Llama3.2-3B, Poison Position=50, k=502025.09 | 34.1 | |
| AstuteRAGModel=Llama3.2-3B, Attack Position=@Pos 50, k=502025.09 | 33.9 |