HiFi-RAG: Hierarchical Content Filtering and Two-Pass Generation for Open-Domain RAG

TL;DR

HiFi-RAG introduces a hierarchical content filtering pipeline for open-domain RAG, replacing standard vector search with a gatekeeping flow using Gemini 2.5 Flash and Gemini 2.5 Pro. The system performs query reformulation, URL filtering, hierarchical content parsing, and a two-pass generation with post-hoc citation verification to improve factuality and attribution. It achieves notable gains on the MMU-RAGent validation set and the post-cutoff Test2025 dataset, illustrating robustness to knowledge up to January 2025 and beyond. The approach balances speed, cost, and accuracy, offering a practical path for up-to-date, reliable open-domain QA.

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.