This paper presents new methods that have the potential to improve privacy process efficiency with LLM and RAG. To reduce hallucination, we continually pre-train the base LLM model with a privacy-specific knowledge base and then augment it with a semantic RAG layer. Our evaluations demonstrate that this approach enhances the model performance (as much as doubled metrics compared to out-of-box LLM) in handling privacy-related queries, by grounding responses with factual information which reduces inaccuracies.
@article{arxiv.2410.02825,
title = {Ingest-And-Ground: Dispelling Hallucinations from Continually-Pretrained LLMs with RAG},
author = {Chenhao Fang and Derek Larson and Shitong Zhu and Sophie Zeng and Wendy Summer and Yanqing Peng and Yuriy Hulovatyy and Rajeev Rao and Gabriel Forgues and Arya Pudota and Alex Goncalves and Hervé Robert},
journal= {arXiv preprint arXiv:2410.02825},
year = {2024}
}