English
Related papers

Related papers: Automatically Identifying Local and Global Circuit…

200 papers

A key goal in mechanistic interpretability is circuit analysis: finding sparse subgraphs of models corresponding to specific behaviors or capabilities. However, MLP sublayers make fine-grained circuit analysis on transformer-based language…

Machine Learning · Computer Science 2024-11-08 Jacob Dunefsky , Philippe Chlenski , Neel Nanda

Decomposing model activations into interpretable components is a key open problem in mechanistic interpretability. Sparse autoencoders (SAEs) are a popular method for decomposing the internal activations of trained transformers into sparse,…

Machine Learning · Computer Science 2024-06-26 Connor Kissane , Robert Krzyzanowski , Joseph Isaac Bloom , Arthur Conmy , Neel Nanda

This paper introduces an efficient and robust method for discovering interpretable circuits in large language models using discrete sparse autoencoders. Our approach addresses key limitations of existing techniques, namely computational…

Computation and Language · Computer Science 2024-05-22 Charles O'Neill , Thang Bui

A widely used strategy to discover and understand language model mechanisms is circuit analysis. A circuit is a minimal subgraph of a model's computation graph that executes a specific task. We identify a gap in existing circuit discovery…

Machine Learning · Computer Science 2025-02-10 Tal Haklay , Hadas Orgad , David Bau , Aaron Mueller , Yonatan Belinkov

Through considerable effort and intuition, several recent works have reverse-engineered nontrivial behaviors of transformer models. This paper systematizes the mechanistic interpretability process they followed. First, researchers choose a…

Machine Learning · Computer Science 2023-10-31 Arthur Conmy , Augustine N. Mavor-Parker , Aengus Lynch , Stefan Heimersheim , Adrià Garriga-Alonso

Mechanistic interpretability seeks to reverse-engineer neural network computations into human-understandable algorithms, yet extracting sparse computational circuits from billion-parameter language models remains challenging due to…

Machine Learning · Computer Science 2026-01-21 Mohammed Mudassir Uddin , Shahnawaz Alam , Mohammed Kaif Pasha

We introduce methods for discovering and applying sparse feature circuits. These are causally implicated subnetworks of human-interpretable features for explaining language model behaviors. Circuits identified in prior work consist of…

Machine Learning · Computer Science 2025-03-28 Samuel Marks , Can Rager , Eric J. Michaud , Yonatan Belinkov , David Bau , Aaron Mueller

Transparency of neural networks' internal reasoning is at the heart of interpretability research, adding to trust, safety, and understanding of these models. The field of mechanistic interpretability has recently focused on studying…

Artificial Intelligence · Computer Science 2026-04-17 Nina Żukowska , Wolfgang Stammer , Bernt Schiele , Jonas Fischer

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

Following their success across many domains, transformers have also proven effective for symbolic regression (SR); however, the internal mechanisms underlying their generation of mathematical operators remain largely unexplored. Although…

Machine Learning · Computer Science 2026-02-04 Arco van Breda , Erman Acar

Neural network models have achieved high performance on a wide variety of complex tasks, but the algorithms that they implement are notoriously difficult to interpret. It is often necessary to hypothesize intermediate variables involved in…

Computation and Language · Computer Science 2025-02-13 Michael A. Lepori , Thomas Serre , Ellie Pavlick

Mechanistic interpretability aims to reverse-engineer transformer computations by identifying causal circuits through activation patching. However, scaling these interventions across diverse prompts and task families produces…

Artificial Intelligence · Computer Science 2026-05-08 Ruben Fernandez-Boullon , David N. Olivieri

Mechanistic interpretability aims to understand neural networks by identifying which learned features mediate specific behaviors. Attribution graphs reveal these feature pathways, but interpreting them requires extensive manual analysis --…

Computation and Language · Computer Science 2025-11-11 Giuseppe Birardi

Sparse autoencoders (SAEs) improve interpretability in multimodal models, but it remains unclear whether SAE features form modular, composable units for reasoning-an assumption underlying many intervention-based steering methods. We test…

Artificial Intelligence · Computer Science 2026-03-27 Yunpeng Zhou

Sparse autoencoders (SAEs) have emerged as a powerful tool for uncovering interpretable features in large language models (LLMs) through the sparse directions they learn. However, the sheer number of extracted directions makes comprehensive…

Computation and Language · Computer Science 2025-11-11 Xinyuan Yan , Shusen Liu , Kowshik Thopalli , Bei Wang

The high-level concepts that a neural network uses to perform computation need not be aligned to individual neurons (Smolensky, 1986). Language model interpretability research has thus turned to techniques such as \textit{sparse…

Computation and Language · Computer Science 2026-02-02 Aryaman Arora , Zhengxuan Wu , Jacob Steinhardt , Sarah Schwettmann

Sparse autoencoders (SAEs) are a popular tool for interpreting large language model activations, but their utility in addressing open questions in interpretability remains unclear. In this work, we demonstrate their effectiveness by using…

Machine Learning · Computer Science 2025-04-21 Dmitrii Kharlapenko , Stepan Shabalin , Fazl Barez , Arthur Conmy , Neel Nanda

Much of mechanistic interpretability has focused on understanding the activation spaces of large neural networks. However, activation space-based approaches reveal little about the underlying circuitry used to compute features. To better…

Machine Learning · Computer Science 2025-04-02 Brianna Chrisman , Lucius Bushnaq , Lee Sharkey

Transformer-based models have become state-of-the-art tools in various machine learning tasks, including time series classification, yet their complexity makes understanding their internal decision-making challenging. Existing…

Machine Learning · Computer Science 2025-11-27 Matīss Kalnāre , Sofoklis Kitharidis , Thomas Bäck , Niki van Stein

Sparse autoencoders (SAEs) extract millions of interpretable features from a language model, but flat feature inventories aren't very useful on their own. Domain concepts get mixed with generic and weakly grounded features, while related…

Artificial Intelligence · Computer Science 2026-04-29 John Winnicki , Abeynaya Gnanasekaran , Eric Darve
‹ Prev 1 2 3 10 Next ›