arXiv 2307.078702023-11-07cs.CLcs.AIcs.LG

作为文化视角叠加态的大语言模型

Large Language Models as Superpositions of Cultural Perspectives

Flowers Team, INRIA（法国国家信息与自动化研究所 Flowers 团队）；Graduate School of Information Science and Technology, The University of Tokyo（东京大学信息科学与技术研究生院）；Ubisoft La Forge（育碧 La Forge）；MIT（麻省理工学院）；INSERM UMR1093-CAPS, Université Bourgogne（法国国家健康与医学研究院 UMR1093-CAPS，勃艮第大学）；Robot Cognition Laboratory, Institute Marey（马雷研究所机器人认知实验室）

Grgur Kovač, Masataka Sawayama, Rémy Portelas, Cédric Colas, Peter Ford Dominey, Pierre-Yves Oudeyer

备注：: Preprint

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

该研究提出大语言模型是不同文化视角的叠加态，其价值观和人格特质随语境变化，通过心理学问卷验证了多模型的视角可控性、诱导方法效果及驱动平滑性，为理解LLM特性提供新视角。

摘要：

大语言模型（LLM）常被误导性地认为具有某种人格或一套价值观。我们提出，LLM可被视为具有不同价值观和人格特质的多种视角的叠加态。LLM展现出的价值观和人格特质具有语境依赖性，会随被诱导出的视角而变化（这与人类不同，人类的价值观和人格特质在不同语境下通常更具一致性）。我们提出了视角可控性的概念，指模型能够采纳带有不同价值观和人格特质的多种视角的能力。在实验中，我们使用心理学领域的问卷（PVQ、VSM、IPIP）来研究LLM展现出的价值观和人格特质如何随不同视角变化。通过定性实验，我们发现当提示中（隐含或明确地）暗示不同价值观时，LLM会表达出不同的价值观；即便提示中没有明显暗示相关价值观，LLM仍会表现出不同的价值观，这证明了其语境依赖的特性。随后我们开展了定量实验，研究了不同模型（GPT-4、GPT-3.5、OpenAssistant、StableVicuna、StableLM）的可控性、多种视角诱导方法的有效性，以及模型可驱动性的平滑度。最后我们探讨了本研究更广泛的意义，并列出了一系列相关的科学问题。项目官网为：https://sites.google.com/view/llm-superpositions 。

英文摘要：

Large Language Models (LLMs) are often misleadingly recognized as having a personality or a set of values. We argue that an LLM can be seen as a superposition of perspectives with different values and personality traits. LLMs exhibit context-dependent values and personality traits that change based on the induced perspective (as opposed to humans, who tend to have more coherent values and personality traits across contexts). We introduce the concept of perspective controllability, which refers to a model's affordance to adopt various perspectives with differing values and personality traits. In our experiments, we use questionnaires from psychology (PVQ, VSM, IPIP) to study how exhibited values and personality traits change based on different perspectives. Through qualitative experiments, we show that LLMs express different values when those are (implicitly or explicitly) implied in the prompt, and that LLMs express different values even when those are not obviously implied (demonstrating their context-dependent nature). We then conduct quantitative experiments to study the controllability of different models (GPT-4, GPT-3.5, OpenAssistant, StableVicuna, StableLM), the effectiveness of various methods for inducing perspectives, and the smoothness of the models' drivability. We conclude by examining the broader implications of our work and outline a variety of associated scientific questions. The project website is available at https://sites.google.com/view/llm-superpositions .

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。