arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2602.06103cs.LG
收藏

从单个文档构建忠实且完整的答案

Toward Faithful and Complete Answer Construction from a Single Document

Department of Mechanical Engineering(机械工程系);Iowa State University(爱荷华州立大学)

Zhaoyang Chen, Cody Fleming

AI总结:

EVE通过结构化框架提升文档基础推理的完整性和准确性,同时解决单次生成中覆盖与准确性的权衡问题。

摘要:

现代大型语言模型(LLMs)是通过统计性下一个词预测驱动的强大生成器。尽管在生成流畅文本方面有效,但这种设计使模型偏向于高概率的延续而非基于源内容的完整且忠实的答案。因此,直接应用LLMs缺乏系统机制来确保完整性和忠实性,这从根本上与核心AI安全原则相冲突。为了解决这一限制,我们提出了EVE,一个用于文档基础推理的结构化框架。与自由形式提示不同,EVE将生成限制在结构化、可验证的流程中,将高严谨性推理分解为提取、验证和列举。从经验上,这种设计使召回率、精确率和F1分数同时得到改进:召回率和精确率分别增加24%和29%,F1分数相应提高31%。这有效打破了单次生成中通常存在的覆盖与准确性之间的长期权衡,同时缓解了由于长度限制导致的生成截断问题。同时,我们强调EVE由于自然语言本身的固有歧义性而表现出性能饱和,反映了语言基础推理的根本限制。

英文摘要:

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.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑