"ProvenanceGuard: RAG Retrieval Poisoning Defense Evaluation Dataset"
Abstract
"This dataset supports the paper \"Provenance-Anchored Retrieval Integrity: An Information-Theoretic Framework for Defending RAG Systems Against Retrieval Poisoning\" published in IEEE Transactions on Information Forensics and Security (2026). The dataset contains 120,000 documents across three high-stakes domains: medical (50,000 passages from PubMed abstracts), legal (30,000 passages from CourtListener case summaries), and financial (40,000 passages from SEC EDGAR filings). It includes 12,000 adversarial passages generated using four attack strategies: Targeted Factual Poisoning (TFP), Semantic Camouflage Poisoning (SCP), Ranking Manipulation (RM), and Multi-passage Coordinated Poisoning (MCP). Additionally, the dataset provides 12,000 evaluation queries with ground truth answers and pre-computed provenance index samples for reproducibility. All source data is drawn from public domain repositories. The dataset enables evaluation of retrieval poisoning defenses for Retrieval-Augmented Generation systems."
Read the paper
Related studies
- Knowledge Management Perspective of Generative Artificial Intelligence
- Navigating Generative Artificial Intelligence Promises and Perils for Knowledge and Creative Work
- Characterizing the evolving research on enterprise content management
- Large language models and multimodal foundation models for precision oncology
- From Generative AI to Extended Reality: Multidisciplinary Perspectives on the Challenges, Opportunities, and Future of Educational Computing
- Generative Artificial Intelligence: A New Engine for Advancing Environmental Science and Engineering
Go deeper in Evidence Hub
Evidence Hub adds a study data sheet (design, sample size, evidence level, key numbers), an AI summary and a chat grounded in the paper's text.