Improving Interpretability in Large Language Models through Layer Wise Relevance Propagation for Identifying Latent Reasoning Biases
International Journal of Artificial Intelligence Research · 19.05.2026 · 1 (2)
DOI: 10.66280/ijair.v1i2.142 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Artificial Intelligence Research, online erschienen am 19.05.2026. Verfasst von Miles Redcliffe und 1 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
The rapid deployment of large language models (LLMs) across critical socio-technical infrastructures has necessitated a paradigm shift in how we understand their internal decision-making processes. As these models transition from experimental prototypes to foundational components of legal, medical, and financial systems, the inherent "black box" nature of transformer architectures presents significant risks regarding latent reasoning biases. This research explores the application of Layer Wise Relevance Propagation (LRP) as a systemic diagnostic tool for enhancing interpretability within high-capacity neural networks. Unlike traditional local explanation methods that often fail to capture the hierarchical dependencies of deep attention mechanisms, LRP offers a robust framework for redistributing output logit scores back through the network layers to identify which specific neurons and structural components contribute most heavily to biased outcomes. This paper examines the structural trade-offs between model performance and transparency, proposing a governance framework that integrates LRP-based auditing into the deployment pipeline. We analyze how latent biases are encoded within the model’s embedding spaces and reinforced through multi-head attention layers, arguing that interpretability is not merely a technical luxury but a functional requirement for systemic robustness. By mapping the relevance flow across diverse transformer blocks, this study identifies critical architectural vulnerabilities where bias becomes entrenched. The findings suggest that systemic interventions at the layer level, rather than simple output filtering, are essential for ensuring the ethical alignment and long-term sustainability of artificial intelligence systems in complex societal environments.
Zitieren
APA
Redcliffe, M., & Blackwood, M. (2026). Improving Interpretability in Large Language Models through Layer Wise Relevance Propagation for Identifying Latent Reasoning Biases. International Journal of Artificial Intelligence Research, 1(2). https://doi.org/10.66280/ijair.v1i2.142
Vancouver
Redcliffe M, Blackwood M. Improving Interpretability in Large Language Models through Layer Wise Relevance Propagation for Identifying Latent Reasoning Biases. International Journal of Artificial Intelligence Research. 2026;1(2). doi:10.66280/ijair.v1i2.142
BibTeX
@article{Redcliffe2026,
author = {Redcliffe, Miles and Blackwood, Matthew},
title = {Improving Interpretability in Large Language Models through Layer Wise Relevance Propagation for Identifying Latent Reasoning Biases},
journal = {International Journal of Artificial Intelligence Research},
year = {2026},
volume = {1},
number = {2},
doi = {10.66280/ijair.v1i2.142}
}
EUVIMED