arXiv cs.AI· Abdullah Al Kalbani, David M. Schultz, Alexandra K. Anderson-Frey, Zhonghua Zheng, Catherine McGuire·· 11 小时前AI 评分38
大语言模型能胜任科学文献综述吗？一项基于 102 篇论文基准的测试
Can Large Language Models Review the Scientific Literature? A Test Against a 102-Article Benchmark
AI 导读
一项以西班牙羽流天气模式综述及其 102 篇论文语料为基准的测试显示，六个 AI 工具在检索、解读、综合与提出未来研究方向四个环节均无法全面胜任。初始检索中五个工具合计找到 57 篇文献，最佳单个工具找到 42 篇；机构图书馆权限使 Codex 的发现量从 42 篇升至 82 篇，Claude Code 从 15 篇升至 79 篇。
正文 · 原文
Abstract:Large language models (LLMs) are increasingly used to support scientific literature reviews, yet evaluations often examine individual tasks rather than the review process as a whole. We assess six AI tools across four stages: finding relevant articles, interpreting them, synthesizing across them, and proposing future research. The benchmark is a published review of the Spanish plume weather pattern and its 102-article corpus.
In the initial searches, five tools collectively identified 57 articles, with the best individual tool finding 42. Institutional library access increased discovery from 42 to 82 articles for Codex and from 15 to 79 for Claude Code. Given the complete corpus of 102 PDFs, three tools correctly identified all 18 articles discussing the Spanish plume substantially, but incorrectly classified 40% to 79% of the 84 passing mentions as substantial. The same three tools produced syntheses evaluated blindly by two domain experts. Although largely accurate, the syntheses provided limited insight and sometimes contained unsupported claims, quotation errors, and exaggerated statements of consensus. Finally, all three tools proposed research directions broader than the Spanish plume literature, although two approached questions identified in the benchmark review.
No single tool performed best across all tasks. We provide eight recommendations for researchers using AI tools in literature reviews. Our findings suggest that LLMs can assist with discovering and interpreting scientific literature, particularly when institutional access is available, but cannot replace expert judgment in evaluating evidence or synthesizing research.
| Comments: | 34 pages, 7 figures, 6 tables. Data, prompts, analysis scripts and results: this https URL (GitHub: this https URL) |
| Subjects: | Artificial Intelligence (cs.AI); Digital Libraries (cs.DL) |
| MSC classes: | 68T50, 68P20, 68T07 |
| Cite as: | arXiv:2611.05459 [cs.AI] |
| (or arXiv:2611.05459v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2611.05459 arXiv-issued DOI via DataCite |
Submission history
From: Abdullah Al Kalbani [view email]
[v1]
Tue, 3 Nov 2026 18:57:20 UTC (811 KB)
来源：arXiv cs.AI · arxiv.org