@inproceedings{tu-etal-2026-paperaudit,
title = "{P}aper{A}udit-Bench: Benchmarking Error Detection in Research Papers for Critical Automated Peer Review",
author = "Tu, Songjun and
Ma, Yiwen and
Lin, Jiahao and
Zhang, Qichao and
Lan, Xiangyuan and
Li, Junfeng. and
Xu, Nan and
Li, Linjing and
Zhao, Dongbin",
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.474/",
pages = "10874--10912",
ISBN = "979-8-89176-459-0",
abstract = "Large language models (LLMs) can generate fluent peer reviews but often miss sparse, subtle, and document-level scientific errors. We introduce PaperAudit-Bench, a benchmark for evaluating paper-level error detection and error-aware automated review. It comprises two components: (1) PaperAudit-Dataset, 220 high-quality AI/ML papers containing more than 2,200 controlled synthetic errors across eight categories; and (2) PaperAudit-Review, an evidence-aware framework that links structured error detection to review generation. Experiments across models, error categories, paper sections, and analysis depths show that sparse long-context auditing remains challenging for current LLMs. Conditioning review generation on explicit, manuscript-grounded error evidence produces more conservative and discriminative assessments that align more closely with human judgments than representative automated-review baselines. The dataset also supports post-training compact detectors through SFT and RL; Qwen3-14B achieves 55.8{\%} F1 at Detection@1 under a fixed held-out protocol. A PubPeer transfer study and deployment in ScienceOne, a scientific foundation-model platform led by the Chinese Academy of Sciences, further illustrate practical applicability.\par Code and Dataset Page: https://github.com/TU2021/PaperAudit\par ScienceOne Manuscript Review: https://www.scienceone.cn/lit/{\#}/assistant"
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="tu-etal-2026-paperaudit">
<titleInfo>
<title>PaperAudit-Bench: Benchmarking Error Detection in Research Papers for Critical Automated Peer Review</title>
</titleInfo>
<name type="personal">
<namePart type="given">Songjun</namePart>
<namePart type="family">Tu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yiwen</namePart>
<namePart type="family">Ma</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jiahao</namePart>
<namePart type="family">Lin</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Qichao</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xiangyuan</namePart>
<namePart type="family">Lan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Junfeng.</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Nan</namePart>
<namePart type="family">Xu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Linjing</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Dongbin</namePart>
<namePart type="family">Zhao</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>Large language models (LLMs) can generate fluent peer reviews but often miss sparse, subtle, and document-level scientific errors. We introduce PaperAudit-Bench, a benchmark for evaluating paper-level error detection and error-aware automated review. It comprises two components: (1) PaperAudit-Dataset, 220 high-quality AI/ML papers containing more than 2,200 controlled synthetic errors across eight categories; and (2) PaperAudit-Review, an evidence-aware framework that links structured error detection to review generation. Experiments across models, error categories, paper sections, and analysis depths show that sparse long-context auditing remains challenging for current LLMs. Conditioning review generation on explicit, manuscript-grounded error evidence produces more conservative and discriminative assessments that align more closely with human judgments than representative automated-review baselines. The dataset also supports post-training compact detectors through SFT and RL; Qwen3-14B achieves 55.8% F1 at Detection@1 under a fixed held-out protocol. A PubPeer transfer study and deployment in ScienceOne, a scientific foundation-model platform led by the Chinese Academy of Sciences, further illustrate practical applicability.\par Code and Dataset Page: https://github.com/TU2021/PaperAudit\par ScienceOne Manuscript Review: https://www.scienceone.cn/lit/#/assistant</abstract>
<identifier type="citekey">tu-etal-2026-paperaudit</identifier>
<location>
<url>https://aclanthology.org/2026.emnlp-main.474/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>10874</start>
<end>10912</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T PaperAudit-Bench: Benchmarking Error Detection in Research Papers for Critical Automated Peer Review
%A Tu, Songjun
%A Ma, Yiwen
%A Lin, Jiahao
%A Zhang, Qichao
%A Lan, Xiangyuan
%A Li, Junfeng.
%A Xu, Nan
%A Li, Linjing
%A Zhao, Dongbin
%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 tu-etal-2026-paperaudit
%X Large language models (LLMs) can generate fluent peer reviews but often miss sparse, subtle, and document-level scientific errors. We introduce PaperAudit-Bench, a benchmark for evaluating paper-level error detection and error-aware automated review. It comprises two components: (1) PaperAudit-Dataset, 220 high-quality AI/ML papers containing more than 2,200 controlled synthetic errors across eight categories; and (2) PaperAudit-Review, an evidence-aware framework that links structured error detection to review generation. Experiments across models, error categories, paper sections, and analysis depths show that sparse long-context auditing remains challenging for current LLMs. Conditioning review generation on explicit, manuscript-grounded error evidence produces more conservative and discriminative assessments that align more closely with human judgments than representative automated-review baselines. The dataset also supports post-training compact detectors through SFT and RL; Qwen3-14B achieves 55.8% F1 at Detection@1 under a fixed held-out protocol. A PubPeer transfer study and deployment in ScienceOne, a scientific foundation-model platform led by the Chinese Academy of Sciences, further illustrate practical applicability.\par Code and Dataset Page: https://github.com/TU2021/PaperAudit\par ScienceOne Manuscript Review: https://www.scienceone.cn/lit/#/assistant
%U https://aclanthology.org/2026.emnlp-main.474/
%P 10874-10912
Markdown (Informal)
[PaperAudit-Bench: Benchmarking Error Detection in Research Papers for Critical Automated Peer Review](https://aclanthology.org/2026.emnlp-main.474/) (Tu et al., EMNLP 2026)
ACL
- Songjun Tu, Yiwen Ma, Jiahao Lin, Qichao Zhang, Xiangyuan Lan, Junfeng. Li, Nan Xu, Linjing Li, and Dongbin Zhao. 2026. PaperAudit-Bench: Benchmarking Error Detection in Research Papers for Critical Automated Peer Review. In Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing, pages 10874–10912, Budapest, Hungary. Association for Computational Linguistics.