# Page-Aware Retrieval-Augmented Generation for EvalLLM 2026: A Five-Variant Study on French PDFs

arXiv: https://arxiv.org/abs/2609.34776 · HTML: https://mlsanity.com/paper/2609.34776/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in RAG pipeline design and evaluation, one of 15 directions in Retrieval-augmented generation.
- Builds on 9 of its 13 references that are in the map.
- First posted on arXiv 28 Sep 2026.

## Placed in

- Retrieval-augmented generation / [RAG pipeline design and evaluation](https://mlsanity.com/rag/rag-pipeline-design-and-evaluation/)

## Builds on (most cited first, 9 of 9)

- [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://mlsanity.com/paper/2005.11401/) (2020)
- [Dense Passage Retrieval for Open-Domain Question Answering](https://mlsanity.com/paper/2004.04906/) (2020)
- [Interleaving Retrieval with Chain-of-Thought Reasoning for Knowledge-Intensive Multi-Step Questions](https://mlsanity.com/paper/2212.10509/) (2022)
- [Passage Re-ranking with BERT](https://mlsanity.com/paper/1901.04085/) (2019)
- [Enabling Large Language Models to Generate Text with Citations](https://mlsanity.com/paper/2305.14627/) (2023)
- [KILT: a Benchmark for Knowledge Intensive Language Tasks](https://mlsanity.com/paper/2009.02252/) (2020)
- [Multilingual E5 Text Embeddings: A Technical Report](https://mlsanity.com/paper/2402.05672/) (2024)
- [An Analysis of Fusion Functions for Hybrid Retrieval](https://mlsanity.com/paper/2210.11934/) (2022)
- [CODENS: Transforming Code Changes into Living, Accessible, and Queryable Documentation](https://mlsanity.com/paper/2607.18356/) (2026)

## Abstract (arXiv)

We study retrieval-augmented generation (RAG) for questions about French PDF documents when both the answer and its supporting document pages are evaluated. Five system variants add dense retrieval, rank fusion, reranking, and query decomposition to a BM25 baseline. On 595 challenge questions, the complete system scores 0.4450 MRR@10 and 0.4013 Recall@10, compared with 0.3430 and 0.2994 for BM25. Dense retrieval alone and a simple lexical--dense fusion both underperform BM25. Reranking improves the hybrid system, whereas adding query decomposition produces the largest further gain, with higher latency and more detected output artifacts. The complete system slightly exceeds the reported anonymous overall mean on two answer metrics but falls below it on most page-retrieval metrics. These results identify accurate page selection, rather than semantic retrieval in isolation, as the main opportunity for improvement in this setting.