cs.LGNov 10, 2026

Mechanistic Interpretability of Deep Neural Networks: Methods, Frameworks, and Perspectives

Authors: Lorenzo Cirillo, Irene Amerini, Andrea Cavallaro

Abstract

Deep neural networks underpin fundamental tasks in a wide range of signal domains, such as image, audio, and language processing. However, their internal mechanisms are notably difficult to interpret. Internal activations are structured signals transformed across layers that may encode a multitude of interdependent relationships, as individual neurons frequently represent multiple underlying concepts. Mechanistic interpretability aims to uncover the internal mechanisms of neural networks by identifying components whose activations are causally responsible for specific behaviors. In such a context, this article provides an overview of advances in mechanistic interpretability for the analysis of causal inference in deep neural networks.

Figures & tables

Explore similar work

CardsList
  1. Mechanistic Interpretability for Neural Networks: Circuits, Sparse Features and Symbolic Reasoning

    Jul 8, 2026Pranav Sawant, Jakub KrejčíTransformer InterpretabilityMechanistic Interpretability

  2. Tracking Equivalent Mechanistic Interpretations Across Neural Networks

    Mar 31, 2026Alan Sun, Mariya TonevaMechanistic InterpretabilityTransformer Interpretability

  3. Causal Interpretation of Neural Network Computations with Contribution Decomposition

    Mar 6, 2026Joshua Brendan Melander, Zaki Alaoui, Shenghua Liu +2Neural Network InterpretabilityMechanistic Interpretability