arXiv cs.CY· Mina Cho, Russell J. Funk, Alok Gupta, Mochen Yang·· 13 小时前AI 评分52
研究发现大语言模型比专家本人更能解释专家知识
Large Language Models Can Explain Experts Better Than Experts Themselves
AI 导读
arXiv 论文提出，大语言模型可从专家行为中提取难以言说的隐性知识，并支持向新手转移。研究先找出能提升 LLM 外化隐性知识的提示策略，再评估其下游效用，结果显示融入专家行为的 LLM 外化知识提升了决策质量、缩小了专家与新手的分差，表现优于人类专家自述的知识。作者进一步表明 LLM 是在从专家行为中学习并提取知识而非简单模仿回答，且结论在不同 LLM 与信息检索方法上均可复现。
正文 · 原文
Abstract:Capturing tacit knowledge, or the "know-how" embedded in individual experiences and contexts, is a central challenge that limits the preservation and transfer of valuable expertise. Traditional methods rely on experts' ability to articulate their own reasoning, even though much of tacit knowledge is difficult to verbalize. Here, we show that Large Language Models (LLMs) can uncover tacit knowledge from expert behavior and support knowledge transfer from experts to novices. We first identify prompting strategies that improve LLMs' ability to externalize tacit knowledge (i.e., transform tacit into explicit knowledge), and then evaluate the downstream utility of the resulting LLM-externalized tacit knowledge. Our analysis reveals that LLM-externalized tacit knowledge incorporating expert behavior improves decision quality and helps reduce the expert-novice score gap, outperforming knowledge articulated by human experts. We further show that LLMs meaningfully learn and extract knowledge from expert behavior rather than simply mimicking expert responses, and that these findings generalize across different LLMs and information retrieval methods. These findings reaffirm Polanyi's Paradox---that "we can know more than we can tell"---and suggest that LLMs can help overcome human experts' articulation bottleneck that has long challenged tacit knowledge externalization.
| Comments: | Main text (29 pages) with Supplementary Information appended (29 pages). Supersedes v1 |
| Subjects: | Human-Computer Interaction (cs.HC); Computers and Society (cs.CY) |
| Cite as: | arXiv:2608.07488 [cs.HC] |
| (or arXiv:2608.07488v2 [cs.HC] for this version) | |
| https://doi.org/10.48550/arXiv.2608.07488 arXiv-issued DOI via DataCite |
Submission history
From: Russell Funk [view email]
[v1]
Sun, 14 Jun 2026 01:48:25 UTC (691 KB)
[v2]
Wed, 21 Oct 2026 03:22:57 UTC (1,886 KB)
来源：arXiv cs.CY · arxiv.org