{"ID":2824746,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2512.22442","arxiv_id":"2512.22442","title":"HiFi-RAG: Hierarchical Content Filtering and Two-Pass Generation for Open-Domain RAG","abstract":"Retrieval-Augmented Generation (RAG) in open-domain settings faces significant challenges regarding irrelevant information in retrieved documents and the alignment of generated answers with user intent. We present HiFi-RAG (Hierarchical Filtering RAG), the winning closed-source system in the Text-to-Text static evaluation of the MMU-RAGent NeurIPS 2025 Competition. Our approach moves beyond standard embedding-based retrieval via a multi-stage pipeline. We leverage the speed and cost-efficiency of Gemini 2.5 Flash (4-6x cheaper than Pro) for query formulation, hierarchical content filtering, and citation attribution, while reserving the reasoning capabilities of Gemini 2.5 Pro for final answer generation. On the MMU-RAGent validation set, our system outperformed the baseline, improving ROUGE-L to 0.274 (+19.6%) and DeBERTaScore to 0.677 (+6.2%). On Test2025, our custom dataset evaluating questions that require post-cutoff knowledge (post January 2025), HiFi-RAG outperforms the parametric baseline by 57.4% in ROUGE-L and 14.9% in DeBERTaScore.","short_abstract":"Retrieval-Augmented Generation (RAG) in open-domain settings faces significant challenges regarding irrelevant information in retrieved documents and the alignment of generated answers with user intent. We present HiFi-RAG (Hierarchical Filtering RAG), the winning closed-source system in the Text-to-Text static evaluat...","url_abs":"https://arxiv.org/abs/2512.22442","url_pdf":"https://arxiv.org/pdf/2512.22442v1","authors":"[\"Cattalyya Nuengsigkapian\"]","published":"2025-12-27T02:37:40Z","proceeding":"cs.CL","tasks":"[\"cs.CL\",\"cs.AI\",\"cs.IR\",\"cs.LG\"]","methods":"[\"RAG\"]","has_code":false}