New RAG Paper

[Trae1ounG]

Hi, we propose a new RAG paradigm which cooperate parametric knowledge with contextual knowledge, improving performance while reducing RAG hallucination.
Paper link: https://arxiv.org/abs/2503.23895