Exploring the Potential of Large Language Models for Text-Based Personality Prediction

This paper explores the potential of large language models (LLMs) for text-based personality prediction, aligning with the Big Five theory. Our study was deliberately confined to a single dimension of Big Five personality traits, namely extraversion, in order to conduct an in-depth exploration of the potential of LLMs for such tasks. The study involves using advanced prompt engineering techniques to influence model behavior, developing methods to effectively describe personality traits, and assessing the impact of prompt modifications on model performance. We provide a comparative analysis of two approaches to personality prediction: Prompt Engineering and traditional methods, highlighting a novel direction in the field.