Latent Structure of Affective Representations in Large Language Models

TL;DR

This paper investigates the geometric structure of affective representations in large language models, revealing alignment with psychological models, linear approximability, and utility for uncertainty quantification.

Contribution

It demonstrates that LLMs learn affective representations consistent with psychological models, supporting linearity assumptions and enabling uncertainty measurement.

Findings

LLMs learn affective representations aligned with valence-arousal models.

Representations exhibit nonlinear but linearly approximable geometric structure.

Latent space can be used to quantify uncertainty in emotion tasks.

Abstract

The geometric structure of latent representations in large language models (LLMs) is an active area of research, driven in part by its implications for model transparency and AI safety. Existing literature has focused mainly on general geometric and topological properties of the learnt representations, but due to a lack of ground-truth latent geometry, validating the findings of such approaches is challenging. Emotion processing provides an intriguing testbed for probing representational geometry, as emotions exhibit both categorical organization and continuous affective dimensions, which are well-established in the psychology literature. Moreover, understanding such representations carries safety relevance. In this work, we investigate the latent structure of affective representations in LLMs using geometric data analysis tools. We present three main findings. First, we show that LLMs learn coherent latent representations of affective emotions that align with widely used valence--arousal models from psychology. Second, we find that these representations exhibit nonlinear geometric structure that can nonetheless be well-approximated linearly, providing empirical support for the linear representation hypothesis commonly assumed in model transparency methods. Third, we demonstrate that the learned latent representation space can be leveraged to quantify uncertainty in emotion processing tasks. Our findings suggest that LLMs acquire affective representations with geometric structure paralleling established models of human emotion, with practical implications for model interpretability and safety.