English
Related papers

Related papers: How does GPT-2 Predict Acronyms? Extracting and Un…

200 papers

Mechanistic interpretability aims to reverse engineer neural networks by uncovering which high-level algorithms they implement. Causal abstraction provides a precise notion of when a network implements an algorithm, i.e., a causal model of…

Machine Learning · Computer Science 2025-03-17 Theodora-Mara Pîslar , Sara Magliacane , Atticus Geiger

Pretrained transformer-based Language Models (LMs) are well-known for their ability to achieve significant improvement on text classification tasks with their powerful word embeddings, but their black-box nature, which leads to a lack of…

Computation and Language · Computer Science 2024-12-23 Ximing Wen , Wenjuan Tan , Rosina O. Weber

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

Recent advances in language model interpretability have identified circuits, critical subnetworks that replicate model behaviors, yet how knowledge is structured within these crucial subnetworks remains opaque. To gain an understanding…

Computation and Language · Computer Science 2025-07-17 Huaizhi Ge , Frank Rudzicz , Zining Zhu

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

Recent advances in interpretability suggest we can project weights and hidden states of transformer-based language models (LMs) to their vocabulary, a transformation that makes them more human interpretable. In this paper, we investigate LM…

Computation and Language · Computer Science 2023-11-27 Shahar Katz , Yonatan Belinkov

To produce accurate predictions, language models (LMs) must balance between generalization and memorization. Yet, little is known about the mechanism by which transformer LMs employ their memorization capacity. When does a model decide to…

Computation and Language · Computer Science 2023-02-14 Adi Haviv , Ido Cohen , Jacob Gidron , Roei Schuster , Yoav Goldberg , Mor Geva

The recent field of neural algorithmic reasoning (NAR) studies the ability of graph neural networks (GNNs) to emulate classical algorithms like Bellman-Ford, a phenomenon known as algorithmic alignment. At the same time, recent advances in…

Machine Learning · Computer Science 2026-02-26 Jesse He , Helen Jenne , Max Vargas , Davis Brown , Gal Mishne , Yusu Wang , Henry Kvinge

Transformer-based language models exhibit complex and distributed behavior, yet their internal computations remain poorly understood. Existing mechanistic interpretability methods typically treat attention heads and multilayer perceptron…

Machine Learning · Computer Science 2025-11-26 Areeb Ahmad , Abhinav Joshi , Ashutosh Modi

While next-token prediction (NTP) has been the standard objective for training language models, it often struggles to capture global structure in reasoning tasks. Multi-token prediction (MTP) has recently emerged as a promising alternative,…

Machine Learning · Computer Science 2026-04-15 Jianhao Huang , Zhanpeng Zhou , Renqiu Xia , Baharan Mirzasoleiman , Weijie Su , Wei Huang

Mechanistic interpretability work attempts to reverse engineer the learned algorithms present inside neural networks. One focus of this work has been to discover 'circuits' -- subgraphs of the full model that explain behaviour on specific…

Machine Learning · Computer Science 2024-07-12 Joseph Miller , Bilal Chughtai , William Saunders

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

Mechanistic interpretability aims to explain neural model behaviour by reverse-engineering learned computational structure into human-understandable components. Without a formal framework, however, mechanistic explanations cannot be…

Machine Learning · Computer Science 2026-05-12 Ward Gauderis , Thomas Dooms , Steven T. Holmer , Kola Ayonrinde , Geraint A. Wiggins

A fundamental question in interpretability research is to what extent neural networks, particularly language models, implement reusable functions through subnetworks that can be composed to perform more complex tasks. Recent advances in…

Machine Learning · Computer Science 2025-06-24 Philipp Mondorf , Sondre Wold , Barbara Plank

Pre-trained language models can be surprisingly adept at tasks they were not explicitly trained on, but how they implement these capabilities is poorly understood. In this paper, we investigate the basic mathematical abilities often…

Computation and Language · Computer Science 2023-11-03 Michael Hanna , Ollie Liu , Alexandre Variengien

Interpretable insights from predictive models remain critical in bio-statistics, particularly when assessing causality, where classical statistical and machine learning methods often provide inherent clarity. While Neural Networks (NNs)…

Applications · Statistics 2025-05-02 Jean-Baptiste A. Conan

Reliance on spurious correlations (shortcuts) has been shown to underlie many of the successes of language models. Previous work focused on identifying the input elements that impact prediction. We investigate how shortcuts are actually…

Machine Learning · Computer Science 2025-05-12 Leon Eshuijs , Shihan Wang , Antske Fokkens

Architectural obfuscation - e.g., permuting hidden-state tensors, linearly transforming embedding tables, or remapping tokens - has recently gained traction as a lightweight substitute for heavyweight cryptography in privacy-preserving…

Cryptography and Security · Computer Science 2025-06-24 Marcos Florencio , Thomas Barton

Graphical tensor notation is a simple way of denoting linear operations on tensors, originating from physics. Modern deep learning consists almost entirely of operations on or between tensors, so easily understanding tensor operations is…

Machine Learning · Computer Science 2024-02-06 Jordan K. Taylor

Although it is known that transformer language models (LMs) pass features from early layers to later layers, it is not well understood how this information is represented and routed by the model. We analyze a mechanism used in two LMs to…

Computation and Language · Computer Science 2025-05-12 Jack Merullo , Carsten Eickhoff , Ellie Pavlick