跳到正文
研踪PaperTrail
← 研踪 PaperTrail

arXiv:2611.23530 · v1

Dreaming Up Authors: Factual Prevalence and LLM Hallucinations in Academic Authorship Attribution

首次提交 2026-11-18 05:37
版本更新 2026-11-18 05:37
雷达观测 2026-11-19 19:08
arXiv ↗PDF ↗

摘要

Large Language Models (LLMs) are widely used for question answering but are prone to hallucinations, yet the underlying causes, particularly the role of factual prevalence in training data, remain underexplored. In this work, we study this through academic authorship attribution, hypothesizing that hallucinations are more likely for less prevalent papers. Using citation count as a proxy for factual prevalence, we curate a dataset of 1.92 million papers across eight scientific disciplines with stratified samples by citation level. Experiments on three state-of-the-art LLMs confirm our hypothesis, revealing a robust inverse relationship between hallucination rate and citation count across all models and disciplines, with hallucination rates exceeding 98% for rarely cited papers and falling below 37% for highly cited ones. Our in-depth analysis further shows that these failures primarily reflect retrieval limitations rather than knowledge absence, and hidden-state probing of open-source models shows that factual prevalence is encoded in pre-generation representations, providing convergent mechanistic evidence for the prevalence-hallucination relationship. The code and data are available at https://github.com/quinnzhong1/dreaming-up-authors.

作者备注:Published in EMNLP 2026 (Main Conference). 37 pages. Code and data: https://github.com/quinnzhong1/dreaming-up-authors. Anthology: https://aclanthology.org/2026.emnlp-main.2372/

期刊信息:Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing (EMNLP 2026), pages 53827-53863, Budapest, Hungary

AI 论文解读

基于摘要 · 非全文审阅

此版本尚未生成 AI 解读。解读由后台任务生成并按论文版本保存,登录后共享已有结果。

版本差异

目前只收录了一个版本。收录下一版后,可以对照标题、作者、分类和摘要。

观测到的版本

相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。

v1 · 2026-11-18 05:37 · 最新观测版本

Dreaming Up Authors: Factual Prevalence and LLM Hallucinations in Academic Authorship Attribution

Large Language Models (LLMs) are widely used for question answering but are prone to hallucinations, yet the underlying causes, particularly the role of factual prevalence in training data, remain underexplored. In this work, we study this through academic authorship attribution, hypothesizing that hallucinations are more likely for less prevalent papers. Using citation count as a proxy for factual prevalence, we curate a dataset of 1.92 million papers across eight scientific disciplines with stratified samples by citation level. Experiments on three state-of-the-art LLMs confirm our hypothesis, revealing a robust inverse relationship between hallucination rate and citation count across all models and disciplines, with hallucination rates exceeding 98% for rarely cited papers and falling below 37% for highly cited ones. Our in-depth analysis further shows that these failures primarily reflect retrieval limitations rather than knowledge absence, and hidden-state probing of open-source models shows that factual prevalence is encoded in pre-generation representations, providing convergent mechanistic evidence for the prevalence-hallucination relationship. The code and data are available at https://github.com/quinnzhong1/dreaming-up-authors.

查看此版本解读 →查看此版本原文 ↗