Hidden Secrets in the arXiv: Discovering, Analyzing, and Preventing Unintentional Information Disclosure in Source Files of Scientific Preprints

2026/04/22 by Jan Pennekamp, Johannes Lohmöller, David Schütte +3 · 7 voices
Computer Science · Decision Sciences · #Academic Publishing and Open Access #Open Source Software Innovations #Research Data Management Practices #cs.CR

paper · pdf · doi:10.1109/sp63933.2026.00217

arxiv published 2026/04/22 · arxiv updated 2026/04/22 · openalex created_date 2026/05/13 · openalex publication_date 2026/05/18 · openalex updated_date 2026/07/29

Abstract

3891

Citations

• You Have Been LaTeXpOsEd: A Systematic Analysis of Information Leakage in Preprint Archives Using Large Language Models
• AblationBench: Evaluating Automated Planning of Ablations in Empirical AI Research
• Modular versus Hierarchical: A Structural Signature of Topic Popularity in Mathematical Research
• WirelessMathBench: A Mathematical Modeling Benchmark for LLMs in Wireless Communications
• Writing Patterns Reveal a Hidden Division of Labor in Scientific Teams
• Text Classification in the LLM Era -- Where do we stand?
• Hidden Division of Labor in Scientific Teams Revealed Through 1.6 Million LaTeX Files
• WithdrarXiv: A Large-Scale Dataset for Retraction Study
• Text Clustering as Classification with LLMs
• LATTE: Improving Latex Recognition for Tables and Formulae with Iterative Refinement
• Tele-LLMs: A Series of Specialized Large Language Models for Telecommunications
• Language Ranker: A Metric for Quantifying LLM Performance Across High and Low-Resource Languages
• Quantitative Analysis of AI-Generated Texts in Academic Research: A Study of AI Presence in Arxiv Submissions using AI Detection Tool
• Text mining arXiv: a look through quantitative finance papers
• Secrets Revealed in Container Images: An Internet-wide Study on Occurrence and Impact
• A Comparative Study of Software Secrets Reporting by Secret Detection Tools
• Sparks of Artificial General Intelligence: Early experiments with GPT-4
• SecretBench: A Dataset of Software Secrets
• A framework for improving the accessibility of research papers on arXiv.org
• Exploitation and Sanitization of Hidden Data in PDF Files
• Is preprint the future of science? A thirty year journey of online preprint services
• Can You Accept LaTeX Files from Strangers? Ten Years Later
• A Guide to Posting and Managing Preprints
• DocBank: A Benchmark Dataset for Document Layout Analysis
• Quantifying the Carbon Emissions of Machine Learning
• On the Use of ArXiv as a Dataset
• Scienceography: the study of how science is written
• Positional Effects on Citation and Readership in arXiv

Discussions

• Secrets in the ArXiv [hn, 3 points, 0 comments]
• Looking more into the paper and its breadth of methods. They used an LLM to classify comments, and REGEX to identify interesting comment features e.g. profanity (101 papers of 2.7M in the dataset with [bsky, 2 points, 1 comments]
• （論文はこれ） Hidden Secrets in the arXiv: Discovering, Analyzing, and Preventing Unintentional Information Disclosure in Source Files of Scientific Preprints arxiv.org/abs/2604.20927 [bsky, 1 points, 0 comments]
• Hidden Secrets in the ArXiv: Information Disclosure in Preprint Source Files [hn, 1 points, 0 comments]
• "Hidden Secrets in the arXiv: Discovering, Analyzing, and Preventing Unintentional Information Disclosure in Source Files of Scientific Preprints" Our analysis reveals that nearly every arXiv submissi [mastodon, 0 points, 0 comments]
• Hidden Secrets in the arXiv: Discovering, Analyzing, and Preventing Unintentional Information Disclosure in Source Files of Scientific Preprints https://arxiv.org/abs/2604.20927v1 [bsky, 0 points, 0 comments]
• Paper: arxiv.org/abs/2604.20927 Executive Summary: arxiv.comsys.rwth-aachen.de ALC-NG: github.com/COMSYS/ALC-NG/ [bsky, 0 points, 0 comments]

Related

• You Have Been LaTeXpOsEd: A Systematic Analysis of Information Leakage in Preprint Archives Using Large Language Models
• The arXiv of the future will not look like the arXiv
• ArXiving Before Submission Helps Everyone
• The Rise of GitHub in Scholarly Publications
• Pushed by Accident: A Mixed-Methods Study on Strategies of Handling Secret Information in Source Code Repositories