https://arxiv.org/abs/2312.10794 Skip to main content Cornell University Served from the cloud We gratefully acknowledge support from the Simons Foundation, member institutions, and all contributors. Donate arxiv logo > cs > arXiv:2312.10794 [ ] Help | Advanced Search [All fields ] Search arXiv logo Cornell University Logo [ ] GO quick links * Login * Help Pages * About Computer Science > Machine Learning arXiv:2312.10794 (cs) [Submitted on 17 Dec 2023] Title:A mathematical perspective on Transformers Authors:Borjan Geshkovski, Cyril Letrouit, Yury Polyanskiy, Philippe Rigollet Download a PDF of the paper titled A mathematical perspective on Transformers, by Borjan Geshkovski and 3 other authors Download PDF Abstract:Transformers play a central role in the inner workings of large language models. We develop a mathematical framework for analyzing Transformers based on their interpretation as interacting particle systems, which reveals that clusters emerge in long time. Our study explores the underlying theory and offers new perspectives for mathematicians as well as computer scientists. Subjects: Machine Learning (cs.LG); Analysis of PDEs (math.AP); Dynamical Systems (math.DS) Cite as: arXiv:2312.10794 [cs.LG] (or arXiv:2312.10794v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2312.10794 Focus to learn more arXiv-issued DOI via DataCite Submission history From: Borjan Geshkovski [view email] [v1] Sun, 17 Dec 2023 19:06:29 UTC (7,096 KB) Full-text links: Access Paper: Download a PDF of the paper titled A mathematical perspective on Transformers, by Borjan Geshkovski and 3 other authors * Download PDF * PostScript (view license, view other formats) Current browse context: cs.LG < prev | next > new | recent | 2312 Change to browse by: cs math math.AP math.DS References & Citations * NASA ADS * Google Scholar * Semantic Scholar a export BibTeX citation Loading... BibTeX formatted citation x [loading... ] Data provided by: Bookmark BibSonomy logo Reddit logo (*) Bibliographic Tools Bibliographic and Citation Tools [ ] Bibliographic Explorer Toggle Bibliographic Explorer (What is the Explorer?) [ ] Litmaps Toggle Litmaps (What is Litmaps?) [ ] scite.ai Toggle scite Smart Citations (What are Smart Citations?) ( ) Code, Data, Media Code, Data and Media Associated with this Article [ ] Links to Code Toggle CatalyzeX Code Finder for Papers (What is CatalyzeX?) [ ] DagsHub Toggle DagsHub (What is DagsHub?) [ ] Links to Code Toggle Papers with Code (What is Papers with Code?) [ ] ScienceCast Toggle ScienceCast (What is ScienceCast?) ( ) Demos Demos [ ] Replicate Toggle Replicate (What is Replicate?) [ ] Spaces Toggle Hugging Face Spaces (What is Spaces?) [ ] Spaces Toggle TXYZ.AI (What is TXYZ.AI?) ( ) Related Papers Recommenders and Search Tools [ ] Link to Influence Flower Influence Flower (What are Influence Flowers?) [ ] Connected Papers Toggle Connected Papers (What is Connected Papers?) [ ] Core recommender toggle CORE Recommender (What is CORE?) [ ] IArxiv recommender toggle IArxiv Recommender (What is IArxiv?) * Author * Venue * Institution * Topic ( ) About arXivLabs arXivLabs: experimental projects with community collaborators arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs. Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?) * About * Help * Click here to contact arXiv Contact * Click here to subscribe Subscribe * Copyright * Privacy Policy * Web Accessibility Assistance * arXiv Operational Status Get status notifications via email or slack