Cross-Attention and Encoder-Decoder Transformers: A Logical Characterization

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Ahvonen, Veeti, Heiman, Damian, Kuusisto, Antti, Moreno, Miguel, Selin, Matias
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917473034436608
author Ahvonen, Veeti
Heiman, Damian
Kuusisto, Antti
Moreno, Miguel
Selin, Matias
author_facet Ahvonen, Veeti
Heiman, Damian
Kuusisto, Antti
Moreno, Miguel
Selin, Matias
contents We give a novel logical characterization of encoder-decoder transformers, the foundational architecture for LLMs that also sees use in various settings that benefit from cross-attention. We study such transformers over text in the practical setting of floating-point numbers and soft-attention, characterizing them with a new temporal logic. This logic extends propositional logic with a counting global modality over the encoder input and a past modality over the decoder input. We also give an additional characterization of such transformers via a type of distributed automata, and show that our results are not limited to the specific choices in the architecture and can account for changes in, e.g., masking. Finally, we discuss encoder-decoder transformers in the autoregressive setting.
format Preprint
id arxiv_https___arxiv_org_abs_2605_07705
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Cross-Attention and Encoder-Decoder Transformers: A Logical Characterization
Ahvonen, Veeti
Heiman, Damian
Kuusisto, Antti
Moreno, Miguel
Selin, Matias
Logic in Computer Science
Artificial Intelligence
F.4.1; F.1.1; I.2.0
We give a novel logical characterization of encoder-decoder transformers, the foundational architecture for LLMs that also sees use in various settings that benefit from cross-attention. We study such transformers over text in the practical setting of floating-point numbers and soft-attention, characterizing them with a new temporal logic. This logic extends propositional logic with a counting global modality over the encoder input and a past modality over the decoder input. We also give an additional characterization of such transformers via a type of distributed automata, and show that our results are not limited to the specific choices in the architecture and can account for changes in, e.g., masking. Finally, we discuss encoder-decoder transformers in the autoregressive setting.
title Cross-Attention and Encoder-Decoder Transformers: A Logical Characterization
topic Logic in Computer Science
Artificial Intelligence
F.4.1; F.1.1; I.2.0
url https://arxiv.org/abs/2605.07705