Assessment and manipulation of latent constructs in pre-trained language models using psychometric scales

TL;DR

This paper introduces a method to assess and manipulate latent psychological constructs in language models using reformulated psychometric questionnaires, enabling better interpretability and control of AI behavior.

Contribution

It presents a novel approach to evaluate and modify psychological traits in language models through psychometric scales reformulated as natural language inference prompts.

Findings

Human-like mental health constructs found in models

Psychometric assessment correlates with human psychology theories

Strategies to mitigate undesirable biases in models

Abstract

Human-like personality traits have recently been discovered in large language models, raising the hypothesis that their (known and as yet undiscovered) biases conform with human latent psychological constructs. While large conversational models may be tricked into answering psychometric questionnaires, the latent psychological constructs of thousands of simpler transformers, trained for other tasks, cannot be assessed because appropriate psychometric methods are currently lacking. Here, we show how standard psychological questionnaires can be reformulated into natural language inference prompts, and we provide a code library to support the psychometric assessment of arbitrary models. We demonstrate, using a sample of 88 publicly available models, the existence of human-like mental health-related constructs (including anxiety, depression, and Sense of Coherence) which conform with standard theories in human psychology and show similar correlations and mitigation strategies. The ability to interpret and rectify the performance of language models by using psychological tools can boost the development of more explainable, controllable, and trustworthy models.