PhD Thesis Defence - Raymond Li

Name: Raymond Li

Date: Oct 1st, 2026

Time: 9am

Location: ICCS X836

Supervisor's name: Giuseppe Carenini

Thesis title: Understanding and Improving Language Models through Interpretable Decompositions

Abstract:

The interpretability of a machine learning model refers to the degree to which a human can understand the model's internal structures and reasoning processes. This thesis investigates globally interpretable decompositions of language model representations, breaking opaque representations into human-interpretable components that can be understood and exploited for model improvement. We organize our investigation around three core constructs: structures, the discrete computational patterns that organize how information is aggregated within a model's internals; topics, the latent thematic variables that capture what a model or corpus is about; and features, the monosemantic directions in activation space that disentangle the concepts superimposed in model activations. A comprehensive set of experimental results demonstrates that our methods enhance not only the interpretability of language models, but also their performance, establishing interpretable decompositions as actionable foundations for both understanding and improving these models.