# Rewiring the Transformer with Depth-Wise LSTMs

arXiv: https://arxiv.org/abs/2007.06257 · HTML: https://mlsanity.com/paper/2007.06257/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 13 Jul 2020.

## Built on by (most cited first, 3 of 3)

- [Mega: Moving Average Equipped Gated Attention](https://mlsanity.com/paper/2209.10655/) (2022)
- [Residual Matrix Transformers: Scaling the Size of the Residual Stream](https://mlsanity.com/paper/2506.22696/) (2025)
- [Interpretability in Activation Space Analysis of Transformers: A Focused Survey](https://mlsanity.com/paper/2302.09304/) (2023)

## Abstract (arXiv)

Stacking non-linear layers allows deep neural networks to model complicated functions, and including residual connections in Transformer layers is beneficial for convergence and performance. However, residual connections may make the model "forget" distant layers and fail to fuse information from previous layers effectively. Selectively managing the representation aggregation of Transformer layers may lead to better performance. In this paper, we present a Transformer with depth-wise LSTMs connecting cascading Transformer layers and sub-layers. We show that layer normalization and feed-forward computation within a Transformer layer can be absorbed into depth-wise LSTMs connecting pure Transformer attention layers. Our experiments with the 6-layer Transformer show significant BLEU improvements in both WMT 14 English-German / French tasks and the OPUS-100 many-to-many multilingual NMT task, and our deep Transformer experiments demonstrate the effectiveness of depth-wise LSTM on the convergence and performance of deep Transformers.