Only Pay What You Must Spend: On-Demand Privacy Budget Payment for Differentially Private RAG

arXiv:2609.27406v1 Announce Type: new Abstract: Deploying large language models (LLMs) on sensitive data via Retrieval-Augmented Generation (RAG) introduces severe privacy risks. Recent studies apply Differential Privacy (DP) to LLMs with RAG for formal privacy guarantees. However, existing DP-RAG frameworks rapidly exhaust the privacy budget.

Read the full article at arXiv cs.CR →