CoSy: Evaluating Textual Explanations of Neurons

Laura Kopf, Philine Lou Bommer, Anna Hedström, Sebastian Lapuschkin, Marina MC Höhne, Kirill Bykov

Keywords: Explainable AI, Evaluation of Explainability Methods, Mechanistic Interpretability

TL;DR: We propose CoSy, an automatic evaluation framework for textual explanations of neurons.

Abstract: A crucial aspect of understanding the complex nature of Deep Neural Networks (DNNs) is the ability to explain learned concepts within their latent representations. While methods exist to connect neurons to human-understandable textual descriptions, evaluating the quality of these explanations is challenging due to the lack of a unified quantitative approach. We introduce CoSy (Concept Synthesis), a novel, architecture-agnostic framework for evaluating textual explanations of latent neurons. Given textual explanations, our proposed framework uses a generative model conditioned on textual input to create data points representing the explanations. By comparing the neuron's response to these generated data points and control data points, we can estimate the quality of the explanation. We validate our framework through sanity checks and benchmark various neuron description methods for Computer Vision tasks, revealing significant differences in quality.

Primary Area: Interpretability and explainability

Submission Number: 15125