ProGRank: Probe-Gradient Reranking to Defend Dense-Retriever RAG from Corpus Poisoning

1. 1Identify potential vulnerabilities in RAG models to corpus poisoning
2. 2Implement ProGRank to rerank retrieved evidence using probe-gradient
3. 3Evaluate the effectiveness of ProGRank in defending against corpus poisoning attacks
4. 4Integrate ProGRank into existing RAG systems to improve their security

ML researchers and engineers working on RAG models can benefit from ProGRank to improve the security of their applications, and software engineers can implement this defence in their systems

💡 ProGRank uses probe-gradient reranking to defend Dense-Retriever RAG from corpus poisoning

💡 Defend RAG models from corpus poisoning with ProGRank!