Transformer Architecture · 2020

On Layer Normalization in the Transformer Architecture (pre-norm)

Ruibin Xiong, Yunchang Yang, Di He, et al.

It showed analytically and empirically that placing layer normalization inside the residual branch (Pre-LN) instead of after it (Post-LN) keeps gradients well-scaled at initialization, removing the need for the learning-rate warm-up stage that Post-LN Transformers required for stable training.

Editorial record

Plain-language summary

The paper studies where the layer-normalization step sits relative to the residual connection in a Transformer block. In the original Post-LN design, gradients near the output layer are large at initialization, so training diverges unless you start with a small learning rate and slowly warm it up. By moving the normalization to the input of each sub-layer (Pre-LN), gradients become bounded and roughly uniform across depth, so models train stably without warm-up, tolerate larger learning rates, and converge faster. This made deep Transformers easier and cheaper to train and is why most later large models adopt the Pre-LN arrangement.

Knowledge graph

Relationships

Descendants

Source record

Provenance

Record ID
P-007
Record created
2026-07-13
Last reviewed
2026-07-14
Record version
2

Citation caveat: Citation metadata is approximate and marked unverified in the source dataset.