@article{
gunasekaran2026satformer,
title={{SATF}ormer: Learning When to Use Early Value Residuals in Transformers},
author={Skye Gunasekaran and T{\'e}a Wright and Rui-Jie Zhu and Jason Eshraghian},
journal={Transactions on Machine Learning Research},
issn={2835-8856},
year={2026},
url={https://openreview.net/forum?id=flq3oO3cNr},
note={}
}