Psychometric Evaluation of Large Language Model Embeddings for Personality Trait Prediction.
Journal:
Journal of medical Internet research
Published Date:
Jul 8, 2025
Abstract
BACKGROUND: Recent advancements in large language models (LLMs) have generated significant interest in their potential for assessing psychological constructs, particularly personality traits. While prior research has explored LLMs' capabilities in zero-shot or few-shot personality inference, few studies have systematically evaluated LLM embeddings within a psychometric validity framework or examined their correlations with linguistic and emotional markers. Additionally, the comparative efficacy of LLM embeddings against traditional feature engineering methods remains underexplored, leaving gaps in understanding their scalability and interpretability for computational personality assessment.
