Analysis of mean-field models arising from self-attention dynamics in transformer architectures with layer normalization.

Saved in:
Bibliographic Details
Title: Analysis of mean-field models arising from self-attention dynamics in transformer architectures with layer normalization.
Authors: Burger, Martin1,2 (AUTHOR) martin.burger@desy.de, Kabri, Samira1 (AUTHOR) samira.kabri@desy.de, Korolev, Yury3 (AUTHOR) ymk30@bath.ac.uk, Roith, Tim1 (AUTHOR) tim.roith@desy.de, Weigand, Lukas1 (AUTHOR) lukas.weigand@desy.de
Source: Philosophical Transactions of the Royal Society A: Mathematical, Physical & Engineering Sciences. 6/5/2025, Vol. 383 Issue 2298, p1-48. 48p.
Database: Academic Search Ultimate
Description
ISSN:1364503X
DOI:10.1098/rsta.2024.0233