Contexts are Never Long Enough: Structured Reasoning for Scalable Question Answering over Long Document Sets

TL;DR

SLIDERS is a structured reasoning framework that enables scalable question answering over long document collections by extracting information into a relational database and reasoning with SQL, outperforming existing methods.

Contribution

It introduces a novel approach combining information extraction, data reconciliation, and SQL-based reasoning to handle long documents beyond fixed context windows.

Findings

SLIDERS outperforms all baselines on three long-context benchmarks.

It exceeds GPT-4.1 by 6.6 points on average.

It significantly improves performance on two large-scale new benchmarks.

Abstract

Real-world document question answering is challenging. Analysts must synthesize evidence across multiple documents and different parts of each document. However, any fixed LLM context window can be exceeded as document collections grow. A common workaround is to decompose documents into chunks and assemble answers from chunk-level outputs, but this introduces an aggregation bottleneck: as the number of chunks grows, systems must still combine and reason over an increasingly large body of extracted evidence. We present SLIDERS, a framework for question answering over long document collections through structured reasoning. SLIDERS extracts salient information into a relational database, enabling scalable reasoning over persistent structured state via SQL rather than concatenated text. To make this locally extracted representation globally coherent, SLIDERS introduces a data reconciliation stage that leverages provenance, extraction rationales, and metadata to detect and repair duplicated, inconsistent, and incomplete records. SLIDERS outperforms all baselines on three existing long-context benchmarks, despite all of them fitting within the context window of strong base LLMs, exceeding GPT-4.1 by 6.6 points on average. It also improves over the next best baseline by ~19 and ~32 points on two new benchmarks at 3.9M and 36M tokens, respectively.