Toward Faithful and Complete Answer Construction from a Single Document

2026-02-05

AI Summary

This work addresses the challenge that large language models struggle to simultaneously ensure answer completeness—avoiding omissions—and faithfulness—avoiding hallucinations—in single-document question answering, thereby violating AI safety principles. To overcome this, the authors propose EVE, a structured and verifiable three-stage pipeline comprising information extraction, fact verification, and systematic enumeration, which constrains the generation process and replaces conventional free-form prompting. EVE is the first method to concurrently improve both coverage and accuracy within a single generation pass, breaking the traditional trade-off between these objectives and alleviating output length truncation issues. Experimental results demonstrate that EVE significantly outperforms baseline approaches across multiple metrics, achieving up to 24% higher recall, 29% higher precision, and a 31% improvement in F1 score.

Technology Category

• Natural Language Processing: Question Answering
• Machine Learning: Large Multimodal Models (LMMs)
• Philosophy and Ethics of AI: Safety, Robustness & Trustworthiness

Application Category

• Search and Retrieval-Augmented AI: Large language models for search
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Modern large language models (LLMs) are powerful generators driven by statistical next-token prediction. While effective at producing fluent text, this design biases models toward high-probability continuations rather than exhaustive and faithful answers grounded in source content. As a result, directly applying LLMs lacks systematic mechanisms to ensure both completeness (avoiding omissions) and faithfulness (avoiding unsupported content), which fundamentally conflicts with core AI safety principles. To address this limitation, we present EVE, a structured framework for document-grounded reasoning. Unlike free-form prompting, EVE constrains generation to a structured, verifiable pipeline that decomposes high-rigor reasoning into extraction, validation, and enumeration. Empirically, this design enables consistent and simultaneous improvements in recall, precision, and F1-score: recall and precision increase by up to 24\% and 29\%, respectively, with a corresponding 31\% gain in F1-score. This effectively breaks the long-standing trade-off between coverage and accuracy typical of single-pass LLM generation, while also mitigating generation truncation caused by length limitations. At the same time, we emphasize that EVE exhibits performance saturation due to the inherent ambiguity of natural language, reflecting fundamental limits of language-based reasoning.

Problem

• faithfulness
• completeness
• document-grounded generation
• large language models
• AI safety

Innovation

• structured reasoning
• faithful generation
• completeness
• document-grounded QA
• EVE framework