English
Related papers

Related papers: Towards Understanding and Improving Refusal in Com…

200 papers

Understanding why machine learning models behave the way they do empowers both system designers and end-users in many ways: in model selection, feature engineering, in order to trust and act upon the predictions, and in more intuitive user…

Machine Learning · Statistics 2016-06-20 Marco Tulio Ribeiro , Sameer Singh , Carlos Guestrin

Model compression is a critical area of research in deep learning, in particular in vision, driven by the need to lighten models memory or computational footprints. While numerous methods for model compression have been proposed, most focus…

Machine Learning · Computer Science 2025-04-08 Jeremy Morlier , Mathieu Leonardon , Vincent Gripon

Transformer plays a vital role in the realms of natural language processing (NLP) and computer vision (CV), specially for constructing large language models (LLM) and large vision models (LVM). Model compression methods reduce the memory…

Machine Learning · Computer Science 2024-04-09 Yehui Tang , Yunhe Wang , Jianyuan Guo , Zhijun Tu , Kai Han , Hailin Hu , Dacheng Tao

The development of deep learning algorithms has extensively empowered humanity's task automatization capacity. However, the huge improvement in the performance of these models is highly correlated with their increasing level of complexity,…

Computer Vision and Pattern Recognition · Computer Science 2024-01-19 Eduarda Caldeira , Pedro C. Neto , Marco Huber , Naser Damer , Ana F. Sequeira

We explore the idea of compressing the prompts used to condition language models, and show that compressed prompts can retain a substantive amount of information about the original prompt. For severely compressed prompts, while fine-grained…

Computation and Language · Computer Science 2022-10-10 David Wingate , Mohammad Shoeybi , Taylor Sorensen

The predictive advantage of combining several different predictive models is widely accepted. Particularly in time series forecasting problems, this combination is often dynamic to cope with potential non-stationary sources of variation…

Machine Learning · Statistics 2021-04-06 Vitor Cerqueira , Luis Torgo , Carlos Soares , Albert Bifet

Safety-aligned language models are trained to refuse harmful requests, yet refusal behavior can be suppressed by steering their internal representations. Existing methods do so by ablating a refusal direction from model activations, aiming…

Artificial Intelligence · Computer Science 2026-05-22 Giorgio Piras , Raffaele Mura , Fabio Brau , Maura Pintor , Luca Oneto , Fabio Roli , Battista Biggio

Effective token compression remains a critical challenge for scaling models to handle increasingly complex and diverse datasets. A novel mechanism based on contextual reinforcement is introduced, dynamically adjusting token importance…

Computation and Language · Computer Science 2025-08-11 Naderdel Piero , Zacharias Cromwell , Nathaniel Wainwright , Matthias Nethercott

Current alignment evaluation mostly measures whether models encode dangerous concepts and whether they refuse harmful requests. Both miss the layer where alignment often operates: routing from concept detection to behavioral policy. We…

Machine Learning · Computer Science 2026-05-04 Gregory N. Frank

Recently, interpretable models called self-explaining models (SEMs) have been proposed with the goal of providing interpretability robustness. We evaluate the interpretability robustness of SEMs and show that explanations provided by SEMs…

Machine Learning · Computer Science 2020-07-03 Haizhong Zheng , Earlence Fernandes , Atul Prakash

With the perpetual increase of complexity of the state-of-the-art deep neural networks, it becomes a more and more challenging task to maintain their interpretability. Our work aims to evaluate the effects of adversarial training utilized…

Computer Vision and Pattern Recognition · Computer Science 2023-11-21 Delyan Boychev

Real-world business applications require a trade-off between language model performance and size. We propose a new method for model compression that relies on vocabulary transfer. We evaluate the method on various vertical domains and…

Computation and Language · Computer Science 2024-02-16 Leonidas Gee , Andrea Zugarini , Leonardo Rigutini , Paolo Torroni

Interpretable classification models are built with the purpose of providing a comprehensible description of the decision logic to an external oversight agent. When considered in isolation, a decision tree, a set of classification rules, or…

Machine Learning · Computer Science 2019-03-18 Riccardo Guidotti , Salvatore Ruggieri

Although deep reinforcement learning has become a promising machine learning approach for sequential decision-making problems, it is still not mature enough for high-stake domains such as autonomous driving or medical applications. In such…

Machine Learning · Computer Science 2022-02-25 Claire Glanois , Paul Weng , Matthieu Zimmer , Dong Li , Tianpei Yang , Jianye Hao , Wulong Liu

To deploy machine learning models on-device, practitioners use compression algorithms to shrink and speed up models while maintaining their high-quality output. A critical aspect of compression in practice is model comparison, including…

Human-Computer Interaction · Computer Science 2025-01-27 Angie Boggust , Venkatesh Sivaraman , Yannick Assogba , Donghao Ren , Dominik Moritz , Fred Hohman

Transparency, user trust, and human comprehension are popular ethical motivations for interpretable machine learning. In support of these goals, researchers evaluate model explanation performance using humans and real world applications.…

Artificial Intelligence · Computer Science 2019-10-31 Bernease Herman

Fine-tuning large language models (LLMs) to adapt to evolving safety policies is costly and impractical. Mechanistic interpretability enables inference-time control through latent activation steering, yet its potential for precise,…

Machine Learning · Computer Science 2025-06-06 Shaona Ghosh , Amrita Bhattacharjee , Yftah Ziser , Christopher Parisien

Machine learning models have had discernible achievements in a myriad of applications. However, most of these models are black-boxes, and it is obscure how the decisions are made by them. This makes the models unreliable and untrustworthy.…

Machine Learning · Computer Science 2020-03-23 Raha Moraffah , Mansooreh Karami , Ruocheng Guo , Adrienne Raglin , Huan Liu

Reinforcement learning policies are typically represented by black-box neural networks, which are non-interpretable and not well-suited for safety-critical domains. To address both of these issues, we propose constrained normalizing flow…

Machine Learning · Computer Science 2024-05-03 Finn Rietz , Erik Schaffernicht , Stefan Heinrich , Johannes A. Stork

We propose using mechanistic interpretability -- techniques for reverse engineering model weights into human-interpretable algorithms -- to derive and compactly prove formal guarantees on model performance. We prototype this approach by…

Machine Learning · Computer Science 2024-12-25 Jason Gross , Rajashree Agrawal , Thomas Kwa , Euan Ong , Chun Hei Yip , Alex Gibson , Soufiane Noubir , Lawrence Chan