@inproceedings{bauwens-de-lhoneux-2026-rebpe,
title = "{R}e{BPE}: Iteratively Improving the Internal Structure of a Structured Tokeniser by Mining its Internal Structure",
author = "Bauwens, Thomas and
de Lhoneux, Miryam",
editor = "Demberg, Vera and
Inui, Kentaro and
Marquez, Llu{\'i}s",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {EACL} 2026",
month = mar,
year = "2026",
address = "Rabat, Morocco",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-eacl.211/",
pages = "4075--4090",
ISBN = "979-8-89176-386-9",
abstract = "Recent work has explored pruning merges from BPE subword tokenisers using corpus data as a signal for which merges to prune. We argue that because a BPE tokeniser contains a rich data structure on top of its vocabulary set, this in itself can be used as a guide to modify its merges such that segmentations become more desirable. We apply this argument to one of those pruning algorithms, BPE-knockout, by introducing a new reification step that suggests new merges by inspecting the effects left by pruning. By alternating both processes iteratively until convergence, we get a new BPE tokeniser, ReBPE, which outperforms the original BPE-knockout algorithm on morphological alignment in all 14 languages tested by over 11{\%} F1 on average."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="bauwens-de-lhoneux-2026-rebpe">
<titleInfo>
<title>ReBPE: Iteratively Improving the Internal Structure of a Structured Tokeniser by Mining its Internal Structure</title>
</titleInfo>
<name type="personal">
<namePart type="given">Thomas</namePart>
<namePart type="family">Bauwens</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Miryam</namePart>
<namePart type="family">de Lhoneux</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-03</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: EACL 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Vera</namePart>
<namePart type="family">Demberg</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kentaro</namePart>
<namePart type="family">Inui</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Lluís</namePart>
<namePart type="family">Marquez</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Rabat, Morocco</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-386-9</identifier>
</relatedItem>
<abstract>Recent work has explored pruning merges from BPE subword tokenisers using corpus data as a signal for which merges to prune. We argue that because a BPE tokeniser contains a rich data structure on top of its vocabulary set, this in itself can be used as a guide to modify its merges such that segmentations become more desirable. We apply this argument to one of those pruning algorithms, BPE-knockout, by introducing a new reification step that suggests new merges by inspecting the effects left by pruning. By alternating both processes iteratively until convergence, we get a new BPE tokeniser, ReBPE, which outperforms the original BPE-knockout algorithm on morphological alignment in all 14 languages tested by over 11% F1 on average.</abstract>
<identifier type="citekey">bauwens-de-lhoneux-2026-rebpe</identifier>
<location>
<url>https://aclanthology.org/2026.findings-eacl.211/</url>
</location>
<part>
<date>2026-03</date>
<extent unit="page">
<start>4075</start>
<end>4090</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T ReBPE: Iteratively Improving the Internal Structure of a Structured Tokeniser by Mining its Internal Structure
%A Bauwens, Thomas
%A de Lhoneux, Miryam
%Y Demberg, Vera
%Y Inui, Kentaro
%Y Marquez, Lluís
%S Findings of the Association for Computational Linguistics: EACL 2026
%D 2026
%8 March
%I Association for Computational Linguistics
%C Rabat, Morocco
%@ 979-8-89176-386-9
%F bauwens-de-lhoneux-2026-rebpe
%X Recent work has explored pruning merges from BPE subword tokenisers using corpus data as a signal for which merges to prune. We argue that because a BPE tokeniser contains a rich data structure on top of its vocabulary set, this in itself can be used as a guide to modify its merges such that segmentations become more desirable. We apply this argument to one of those pruning algorithms, BPE-knockout, by introducing a new reification step that suggests new merges by inspecting the effects left by pruning. By alternating both processes iteratively until convergence, we get a new BPE tokeniser, ReBPE, which outperforms the original BPE-knockout algorithm on morphological alignment in all 14 languages tested by over 11% F1 on average.
%U https://aclanthology.org/2026.findings-eacl.211/
%P 4075-4090
Markdown (Informal)
[ReBPE: Iteratively Improving the Internal Structure of a Structured Tokeniser by Mining its Internal Structure](https://aclanthology.org/2026.findings-eacl.211/) (Bauwens & de Lhoneux, Findings 2026)
ACL