A completely uniform transformer for parity

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Kozachinskiy, Alexander, Steifer, Tomasz
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909448558084096
author Kozachinskiy, Alexander
Steifer, Tomasz
author_facet Kozachinskiy, Alexander
Steifer, Tomasz
contents We construct a 3-layer constant-dimension transformer, recognizing the parity language, where neither parameter matrices nor the positional encoding depend on the input length. This improves upon a construction of Chiang and Cholak who use a positional encoding, depending on the input length (but their construction has 2 layers).
format Preprint
id arxiv_https___arxiv_org_abs_2501_02535
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle A completely uniform transformer for parity
Kozachinskiy, Alexander
Steifer, Tomasz
Machine Learning
Artificial Intelligence
We construct a 3-layer constant-dimension transformer, recognizing the parity language, where neither parameter matrices nor the positional encoding depend on the input length. This improves upon a construction of Chiang and Cholak who use a positional encoding, depending on the input length (but their construction has 2 layers).
title A completely uniform transformer for parity
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2501.02535