arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2503.09211cs.LGcs.CL
收藏

为什么LLM无法思考以及如何修复它

Why LLMs Cannot Think and How to Fix It

Institute of Neuro- and Bioinformatics(神经与生物信息学研究所);University of Lübeck(吕贝克大学)

Marius Jahrens, Thomas Martinetz

AI总结:

本文从架构与训练机制论证LLM无法在特征空间中真正思考,并给出支持思维过程的架构修复方向。

摘要:

本文阐明,由于架构约束,当前最先进的大型语言模型(LLM)在根本上无法在特征空间内做出决策或形成“思想”。我们建立了一个涵盖该术语传统理解的“思想”定义,并将其调整以适用于LLM。我们证明,当代LLM的架构设计和语言建模训练方法本质上阻碍了它们参与真正的思维过程。我们主要关注这一理论认识,而非从实验数据中得出的实践洞见。最后,我们提出了在特征空间内实现思维过程的解决方案,并讨论了这些架构修改的更广泛影响。

英文摘要:

This paper elucidates that current state-of-the-art Large Language Models (LLMs) are fundamentally incapable of making decisions or developing "thoughts" within the feature space due to their architectural constraints. We establish a definition of "thought" that encompasses traditional understandings of that term and adapt it for application to LLMs. We demonstrate that the architectural design and language modeling training methodology of contemporary LLMs inherently preclude them from engaging in genuine thought processes. Our primary focus is on this theoretical realization rather than practical insights derived from experimental data. Finally, we propose solutions to enable thought processes within the feature space and discuss the broader implications of these architectural modifications.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑