@inproceedings{xiao-etal-2026-enhancing,
title = "Enhancing Long-Context Reasoning via Verifiable Atomic Skills",
author = "Xiao, Yanling and
Xie, Huaibing and
Zhao, Guoliang and
Dou, Shihan and
Zheng, Nantao and
Wang, Shaolei and
Liu, Yiting and
Zhang, Cheng and
Zhou, Pluto and
Zhang, Zhisong and
Liu, Lemao",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.emnlp-main.1513/",
pages = "34516--34532",
ISBN = "979-8-89176-459-0",
abstract = "Long-context reasoning remains a critical challenge for Large Language Models (LLMs). Current research typically focuses on curating specialized training data by treating long-context reasoning as a monolithic task, yet it is challenging to verify its quality due to its internal cognitive complexity. In this paper, we propose a paradigm shift: viewing long-context reasoning as a hierarchical spectrum of core atomic skills. To operationalize this, we present an Anchor-based Reasoning (AbR) pipeline that automatically synthesizes strictly verifiable training data for each skill, thereby enabling reliable quality. Diagnostic analyses confirm that proficiency in these atomic skills strongly predicts general long-context reasoning performance. By training exclusively on a highly compact set of merely four thousand synthesized samples with Reinforcement Learning (RL), we explicitly sharpen these foundational capabilities. Extensive experiments across six challenging benchmarks demonstrate that this targeted RL causally delivers robust and generalizable performance gains, outperforming DeepSeek-R1-distill-32B by an average margin of 7.7{\%} (improving from 46.3{\%} to 54.0{\%}). Furthermore, in-depth ablations reveal that these skills operate synergistically, establishing a scalable, data-efficient, and verifiable path to advancing long-context reasoning."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="xiao-etal-2026-enhancing">
<titleInfo>
<title>Enhancing Long-Context Reasoning via Verifiable Atomic Skills</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yanling</namePart>
<namePart type="family">Xiao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Huaibing</namePart>
<namePart type="family">Xie</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Guoliang</namePart>
<namePart type="family">Zhao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shihan</namePart>
<namePart type="family">Dou</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Nantao</namePart>
<namePart type="family">Zheng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shaolei</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yiting</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Cheng</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Pluto</namePart>
<namePart type="family">Zhou</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhisong</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Lemao</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hung-Yi</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunipa</namePart>
<namePart type="family">Dev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jessy</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Desmond</namePart>
<namePart type="family">Elliott</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-459-0</identifier>
</relatedItem>
<abstract>Long-context reasoning remains a critical challenge for Large Language Models (LLMs). Current research typically focuses on curating specialized training data by treating long-context reasoning as a monolithic task, yet it is challenging to verify its quality due to its internal cognitive complexity. In this paper, we propose a paradigm shift: viewing long-context reasoning as a hierarchical spectrum of core atomic skills. To operationalize this, we present an Anchor-based Reasoning (AbR) pipeline that automatically synthesizes strictly verifiable training data for each skill, thereby enabling reliable quality. Diagnostic analyses confirm that proficiency in these atomic skills strongly predicts general long-context reasoning performance. By training exclusively on a highly compact set of merely four thousand synthesized samples with Reinforcement Learning (RL), we explicitly sharpen these foundational capabilities. Extensive experiments across six challenging benchmarks demonstrate that this targeted RL causally delivers robust and generalizable performance gains, outperforming DeepSeek-R1-distill-32B by an average margin of 7.7% (improving from 46.3% to 54.0%). Furthermore, in-depth ablations reveal that these skills operate synergistically, establishing a scalable, data-efficient, and verifiable path to advancing long-context reasoning.</abstract>
<identifier type="citekey">xiao-etal-2026-enhancing</identifier>
<location>
<url>https://aclanthology.org/2026.emnlp-main.1513/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>34516</start>
<end>34532</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Enhancing Long-Context Reasoning via Verifiable Atomic Skills
%A Xiao, Yanling
%A Xie, Huaibing
%A Zhao, Guoliang
%A Dou, Shihan
%A Zheng, Nantao
%A Wang, Shaolei
%A Liu, Yiting
%A Zhang, Cheng
%A Zhou, Pluto
%A Zhang, Zhisong
%A Liu, Lemao
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-459-0
%F xiao-etal-2026-enhancing
%X Long-context reasoning remains a critical challenge for Large Language Models (LLMs). Current research typically focuses on curating specialized training data by treating long-context reasoning as a monolithic task, yet it is challenging to verify its quality due to its internal cognitive complexity. In this paper, we propose a paradigm shift: viewing long-context reasoning as a hierarchical spectrum of core atomic skills. To operationalize this, we present an Anchor-based Reasoning (AbR) pipeline that automatically synthesizes strictly verifiable training data for each skill, thereby enabling reliable quality. Diagnostic analyses confirm that proficiency in these atomic skills strongly predicts general long-context reasoning performance. By training exclusively on a highly compact set of merely four thousand synthesized samples with Reinforcement Learning (RL), we explicitly sharpen these foundational capabilities. Extensive experiments across six challenging benchmarks demonstrate that this targeted RL causally delivers robust and generalizable performance gains, outperforming DeepSeek-R1-distill-32B by an average margin of 7.7% (improving from 46.3% to 54.0%). Furthermore, in-depth ablations reveal that these skills operate synergistically, establishing a scalable, data-efficient, and verifiable path to advancing long-context reasoning.
%U https://aclanthology.org/2026.emnlp-main.1513/
%P 34516-34532
Markdown (Informal)
[Enhancing Long-Context Reasoning via Verifiable Atomic Skills](https://aclanthology.org/2026.emnlp-main.1513/) (Xiao et al., EMNLP 2026)
ACL
- Yanling Xiao, Huaibing Xie, Guoliang Zhao, Shihan Dou, Nantao Zheng, Shaolei Wang, Yiting Liu, Cheng Zhang, Pluto Zhou, Zhisong Zhang, and Lemao Liu. 2026. Enhancing Long-Context Reasoning via Verifiable Atomic Skills. In Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing, pages 34516–34532, Budapest, Hungary. Association for Computational Linguistics.