Concept Layers: Enhancing Interpretability and Intervenability via LLM Conceptualization

TL;DR

This paper introduces Concept Layers, a novel method to improve interpretability and intervenability of LLMs by projecting internal representations into an explainable space and automatically selecting relevant concepts, without extensive architectural changes.

Contribution

The work proposes Concept Layers that integrate into existing models, eliminating the need for labeled concept datasets and enabling dynamic user interventions.

Findings

Maintains original model performance and agreement.

Enables meaningful model interventions.

Supports bias mitigation during inference.

Abstract

The opaque nature of Large Language Models (LLMs) has led to significant research efforts aimed at enhancing their interpretability, primarily through post-hoc methods. More recent in-hoc approaches, such as Concept Bottleneck Models (CBMs), offer both interpretability and intervenability by incorporating explicit concept representations. However, these methods suffer from key limitations, including reliance on labeled concept datasets and significant architectural modifications that challenges re-integration into existing system pipelines. In this work, we introduce a new methodology for incorporating interpretability and intervenability into an existing model by integrating Concept Layers (CLs) into its architecture. Our approach projects the model's internal vector representations into a conceptual, explainable vector space before reconstructing and feeding them back into the model. Furthermore, we eliminate the need for a human-selected concept set by algorithmically searching an ontology for a set of concepts that can be either task-specific or task-agnostic. We evaluate CLs across multiple tasks, demonstrating that they maintain the original model's performance and agreement while enabling meaningful interventions. Additionally, we present a proof of concept showcasing an intervenability interface, allowing users to adjust model behavior dynamically, such as mitigating biases during inference.