arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2607.01006cs.CL
收藏

理解大型语言模型

Understanding Large Language Models

Yannik Keller, Thomas Eisenmann

AI总结:

本文综述大型语言模型(LLM)的机制、涌现能力及与人类认知的关系,通过分析注意力机制、符号推理、心智理论等证据,探讨LLM是否真正理解语言,并反对过度简化的人机认知差异观点。

摘要:

大型语言模型(LLM)代表了近年来人工智能和自然语言处理领域最显著的进步之一。然而,关于其机制、能力以及与人类认知的关系,许多紧迫的问题仍然存在高度争议。本章旨在通过讨论关于涌现能力及其在处理层中的机制实现的最新证据,概述我们目前对LLM的理解。我们首先简要概述Transformer架构,强调注意力机制如何使大规模数据集训练成为可能,从而使LLM能够作为通用模型而非专用模型运行。接下来,我们考察LLM的涌现能力,这些能力似乎与人类认知的某些方面相似,包括符号推理、心智理论和欺骗策略。多项研究提供证据表明,LLM能够解决以前被认为需要类人认知的任务。其他研究则揭示了有洞察力的失败案例,阐明了人类与LLM认知之间的差异。伴随这些发现,我们回顾了从神经元激活分析到电路追踪的可解释AI方法。在最后一节中,我们讨论了当前关于LLM真正理解什么与仅仅看起来理解什么的争论。反对AI拟人化的主要论点指出LLM训练目标的简单性,声称LLM的行为更可能是对训练数据的模式记忆而非真正的认知。我们认为这一观点受到关于优化过程和认知能力的误解的引导,并倡导对LLM认知进行更细致的讨论,既不忽视人类与LLM之间的差异,也不通过过于简单的还原论论证排除AI认知的可能性。

英文摘要:

Large Language Models (LLMs) represent one of the most significant advances in AI and natural language processing in recent years. Still, many pressing questions about their mechanisms, capabilities, and relationship to human cognition remain highly debated. This chapter aims to outline our current understanding of LLMs by discussing recent evidence on emerging capabilities and their mechanistic implementation within processing layers. We begin with a concise overview of the Transformer architecture, emphasizing how the attention mechanism enables training on massive datasets, allowing LLMs to function as generalist rather than specialized models. Next, we examine emergent LLM capabilities that appear to resemble aspects of human cognition, including symbolic reasoning, theory of mind, and deception strategies. Several studies provide evidence that LLMs can solve tasks previously thought to require human-like cognition. Other studies reveal insightful failure cases that shed light on the differences between human and LLM cognition. Alongside these findings, we review explainable AI approaches ranging from neuron activation analysis to circuit tracing. In the final section, we address current debates concerning what LLMs genuinely understand versus what they merely appear to understand. Prominent arguments against AI anthropomorphism point to the simplicity of LLM training objectives, claiming that LLM behavior is better explained by pattern memorization of training data than by genuine cognition. We argue that this standpoint is guided by misconceptions about optimization processes and cognitive capacity, and advocate for a more nuanced discussion of LLM cognition that neither dismisses the differences between humans and LLMs nor precludes the possibility of AI cognition through overly simplistic reductionist arguments.

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。

↑