English
Related papers

Related papers: Circuit Insights: Towards Interpretability Beyond …

200 papers

Counting serves as a simple but powerful test of a Large Vision-Language Model's (LVLM's) reasoning; it forces the model to identify each individual object and then add them all up. In this study, we investigate how LVLMs implement counting…

Computer Vision and Pattern Recognition · Computer Science 2026-03-20 Liwei Che , Zhiyu Xue , Yihao Quan , Benlin Liu , Zeru Shi , Michelle Hurst , Jacob Feldman , Ruixiang Tang , Ranjay Krishna , Vladimir Pavlovic

To date, most discoveries of network subcomponents that implement human-interpretable computations in deep vision models have involved close study of single units and large amounts of human labor. We explore scalable methods for extracting…

Computer Vision and Pattern Recognition · Computer Science 2024-04-23 Achyuta Rajaram , Neil Chowdhury , Antonio Torralba , Jacob Andreas , Sarah Schwettmann

Mechanistic interpretability identifies internal circuits responsible for model behaviors, yet translating these findings into human-understandable explanations remains an open problem. We present a pipeline that bridges circuit-level…

Computation and Language · Computer Science 2026-03-12 Ajay Pravin Mahale

The predictive power of neural networks often costs model interpretability. Several techniques have been developed for explaining model outputs in terms of input features; however, it is difficult to translate such interpretations into…

Machine Learning · Computer Science 2017-08-17 Benjamin J. Lengerich , Sandeep Konam , Eric P. Xing , Stephanie Rosenthal , Manuela Veloso

Mechanistic Interpretability (MI) promises a path toward fully understanding how neural networks make their predictions. Prior work demonstrates that even when trained to perform simple arithmetic, models can implement a variety of…

Machine Learning · Computer Science 2024-05-28 Ouail Kitouni , Niklas Nolte , Víctor Samuel Pérez-Díaz , Sokratis Trifinopoulos , Mike Williams

Single neurons in neural networks are often interpretable in that they represent individual, intuitively meaningful features. However, many neurons exhibit $\textit{mixed selectivity}$, i.e., they represent multiple unrelated features. A…

Machine Learning · Statistics 2023-10-19 David Klindt , Sophia Sanborn , Francisco Acosta , Frédéric Poitevin , Nina Miolane

Mechanistic Interpretability (MI) aims to reverse-engineer model behaviors by identifying functional sub-networks. Yet, the scientific validity of these findings depends on their stability. In this work, we argue that circuit discovery is…

Machine Learning · Computer Science 2026-02-04 Maxime Méloux , François Portet , Maxime Peyrard

Automated interpretability pipelines generate natural language descriptions for the concepts represented by features in large language models (LLMs), such as plants or the first word in a sentence. These descriptions are derived using…

Computation and Language · Computer Science 2025-05-30 Yoav Gur-Arieh , Roy Mayan , Chen Agassy , Atticus Geiger , Mor Geva

Mechanistic interpretability seeks to reverse engineer a trained neural network by identifying the minimal subset of internal components. We perform a mechanistic interpretability analysis of the Particle Transformer architecture, trained…

High Energy Physics - Phenomenology · Physics 2026-05-12 Saurabh Rai , Sanmay Ganguly

Automated interpretability systems aim to reduce the need for human labor and scale analysis to increasingly large models and diverse tasks. Recent efforts toward this goal leverage large language models (LLMs) at increasing levels of…

Artificial Intelligence · Computer Science 2026-03-23 Tal Haklay , Nikhil Prakash , Sana Pandey , Antonio Torralba , Aaron Mueller , Jacob Andreas , Tamar Rott Shaham , Yonatan Belinkov

The circuits framework in mechanistic interpretability aims to identify causally important sparse subgraphs of model components, typically evaluated by measuring necessity and sufficiency. We measure circuit reuse, the proportion of…

Computation and Language · Computer Science 2026-05-12 Michael Li , Nishant Subramani

Inner Interpretability is a promising emerging field tasked with uncovering the inner mechanisms of AI systems, though how to develop these mechanistic theories is still much debated. Moreover, recent critiques raise issues that question…

Artificial Intelligence · Computer Science 2024-08-01 Martina G. Vilas , Federico Adolfi , David Poeppel , Gemma Roig

Large language models (LLMs) are often portrayed as merely imitating linguistic patterns without genuine understanding. We argue that recent findings in mechanistic interpretability (MI), the emerging field probing the inner workings of…

Computation and Language · Computer Science 2026-02-26 Pierre Beckmann , Matthieu Queloz

*Automated circuit discovery* is a central tool in mechanistic interpretability for identifying the internal components of neural networks responsible for specific behaviors. While prior methods have made significant progress, they…

Machine Learning · Computer Science 2026-02-20 Itamar Hadad , Guy Katz , Shahaf Bassan

We propose an efficient and interpretable neural network with a novel activation function called the weighted Lehmer transform. This new activation function enables adaptive feature selection and extends to the complex domain, capturing…

Machine Learning · Computer Science 2025-01-28 Masoud Ataei , Xiaogang Wang

Interpreting the inner function of neural networks is crucial for the trustworthy development and deployment of these black-box models. Prior interpretability methods focus on correlation-based measures to attribute model decisions to…

Machine Learning · Computer Science 2023-06-21 Ola Ahmad , Nicolas Bereux , Loïc Baret , Vahid Hashemi , Freddy Lecue

Vision-Language-Action Models (VLAs) have shown remarkable progress towards embodied intelligence. While their architecture partially resembles that of Large Language Models (LLMs), VLAs exhibit higher complexity due to their multi-modal…

Robotics · Computer Science 2026-03-06 Hugo Buurmeijer , Carmen Amo Alonso , Aiden Swann , Marco Pavone

Neural IR architectures, particularly cross-encoders, are highly effective models whose internal mechanisms are mostly unknown. Most works trying to explain their behavior focused on high-level processes (e.g., what in the input influences…

Information Retrieval · Computer Science 2025-07-22 Mathias Vast , Basile Van Cooten , Laure Soulier , Benjamin Piwowarski

The ability of neural networks to represent more features than neurons makes interpreting them challenging. This phenomenon, known as superposition, has spurred efforts to find architectures that are more interpretable than standard…

Machine Learning · Computer Science 2023-05-08 Lee Sharkey

Deep learning continues to play as a powerful state-of-art technique that has achieved extraordinary accuracy levels in various domains of regression and classification tasks, including images, video, signal, and natural language data. The…

Neural and Evolutionary Computing · Computer Science 2022-06-03 Anna Zou , Zhiyuan Li