Saved in:
Bibliographic Details
Main Authors: Cao, Weili, Wang, Jianyou, Zheng, Youze, Bao, Longtian, Zheng, Qirui, Berg-Kirkpatrick, Taylor, Paturi, Ramamohan, Bergen, Leon
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2504.03101
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915434094133248
author Cao, Weili
Wang, Jianyou
Zheng, Youze
Bao, Longtian
Zheng, Qirui
Berg-Kirkpatrick, Taylor
Paturi, Ramamohan
Bergen, Leon
author_facet Cao, Weili
Wang, Jianyou
Zheng, Youze
Bao, Longtian
Zheng, Qirui
Berg-Kirkpatrick, Taylor
Paturi, Ramamohan
Bergen, Leon
contents Handling extremely large documents for question answering is challenging: chunk-based embedding methods often lose track of important global context, while full-context transformers can be prohibitively expensive for hundreds of thousands of tokens. We propose a single-pass document scanning approach that processes the entire text in linear time, preserving global coherence while deciding which sentences are most relevant to the query. On 41 QA benchmarks, our single-pass scanner consistently outperforms chunk-based embedding methods and competes with large language models at a fraction of the computational cost. By conditioning on the entire preceding context without chunk breaks, the method preserves global coherence, which is especially important for long documents. Overall, single-pass document scanning offers a simple solution for question answering over massive text. All code, datasets, and model checkpoints are available at https://github.com/MambaRetriever/MambaRetriever
format Preprint
id arxiv_https___arxiv_org_abs_2504_03101
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Single-Pass Document Scanning for Question Answering
Cao, Weili
Wang, Jianyou
Zheng, Youze
Bao, Longtian
Zheng, Qirui
Berg-Kirkpatrick, Taylor
Paturi, Ramamohan
Bergen, Leon
Computation and Language
Handling extremely large documents for question answering is challenging: chunk-based embedding methods often lose track of important global context, while full-context transformers can be prohibitively expensive for hundreds of thousands of tokens. We propose a single-pass document scanning approach that processes the entire text in linear time, preserving global coherence while deciding which sentences are most relevant to the query. On 41 QA benchmarks, our single-pass scanner consistently outperforms chunk-based embedding methods and competes with large language models at a fraction of the computational cost. By conditioning on the entire preceding context without chunk breaks, the method preserves global coherence, which is especially important for long documents. Overall, single-pass document scanning offers a simple solution for question answering over massive text. All code, datasets, and model checkpoints are available at https://github.com/MambaRetriever/MambaRetriever
title Single-Pass Document Scanning for Question Answering
topic Computation and Language
url https://arxiv.org/abs/2504.03101