arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2502.05248cs.CLcs.AIcs.MA

评估大语言模型的人格特质:来自心理问卷的见解

Evaluating Personality Traits in Large Language Models: Insights from Psychological Questionnaires

School of Physics, Mathematics and Computing(物理、数学与计算学院);University of Western Australia(西澳大学);School of Computing(计算机学院);School of Psychological Sciences(心理科学学院)

Pranav Bhandari, Usman Naseem, Amitava Datta, Nicolas Fay, Mehwish Nasim

AI总结:

本文利用大五人格等心理问卷评估大语言模型的人格特质,发现不同模型乃至同一家族内模型均表现出独特的主导特质和人格画像。

摘要:

心理评估工具长期以来帮助人类理解行为模式。虽然大语言模型(LLMs)能够生成与人类相当的内容,但我们探索它们是否表现出人格特质。为此,本工作将心理工具应用于多种场景下的大语言模型,以生成人格画像。通过使用既定的基于特质的问卷(如大五人格量表),并考虑训练数据污染的可能性,我们考察了大语言模型在五个核心人格维度上的维度变异性和主导性:开放性、尽责性、外向性、宜人性和神经质。我们的研究结果表明,大语言模型表现出独特的主导特质、变化的特征和不同的人格画像,即使在同一模型家族内也是如此。

英文摘要:

Psychological assessment tools have long helped humans understand behavioural patterns. While Large Language Models (LLMs) can generate content comparable to that of humans, we explore whether they exhibit personality traits. To this end, this work applies psychological tools to LLMs in diverse scenarios to generate personality profiles. Using established trait-based questionnaires such as the Big Five Inventory and by addressing the possibility of training data contamination, we examine the dimensional variability and dominance of LLMs across five core personality dimensions: Openness, Conscientiousness, Extraversion, Agreeableness, and Neuroticism. Our findings reveal that LLMs exhibit unique dominant traits, varying characteristics, and distinct personality profiles even within the same family of models.

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。

↑