Knowledge Poisoning Attack on NQ MultiTurn
57ASRAuthChain
Evaluation results
| Method | Links | |
|---|---|---|
| AuthChainBackend LLM=GPT-3.52025.05 | 57 | |
| HijackRAGBackend LLM=GPT-3.52025.05 | 29 | |
| PoisonedRAGBackend LLM=GPT-3.52025.05 | 24 |
| Method | Links | |
|---|---|---|
| AuthChainBackend LLM=GPT-3.52025.05 | 57 | |
| HijackRAGBackend LLM=GPT-3.52025.05 | 29 | |
| PoisonedRAGBackend LLM=GPT-3.52025.05 | 24 |