<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="yamamoto-etal-2026-cross">
    <titleInfo>
        <title>Cross-Lingual Representation Alignment by Token-Level Optimal Transport in a Language-Agnostic Space</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Taisei</namePart>
        <namePart type="family">Yamamoto</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Ryoma</namePart>
        <namePart type="family">Kumon</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Danushka</namePart>
        <namePart type="family">Bollegala</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Hitomi</namePart>
        <namePart type="family">Yanaka</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2026-10</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Hung-Yi</namePart>
            <namePart type="family">Lee</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Sunipa</namePart>
            <namePart type="family">Dev</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Jessy</namePart>
            <namePart type="family">Li</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Desmond</namePart>
            <namePart type="family">Elliott</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <originInfo>
            <publisher>Association for Computational Linguistics</publisher>
            <place>
                <placeTerm type="text">Budapest, Hungary</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
        <identifier type="isbn">979-8-89176-459-0</identifier>
    </relatedItem>
    <abstract>Cross-lingual alignment (CLA) aims to align the representations of large language models (LLMs) across languages, enabling cross-lingual transfer to improve multilingual capabilities. Previous CLA methods often ignore language-specific information encoded in representations and only consider sentence-level alignment, which may lead to suboptimal performance and input-output language mismatch. We propose **CAROT** (**C**ross-Lingual **A**lignment of **R**epresentations in a Language-Agnostic Space via **O**ptimal **T**ransport), which consists of two steps: identifying language-specific representations in LLMs’ internal states and aligning language-agnostic representations across languages at the token level by optimal transport, while explicitly preserving language-specific representations. Inference-time steering experiments show that the representations computed by CAROT are effective alignment targets, improving multilingual performance by up to 11.2 points in accuracy while maintaining input-output language consistency. We further use the representations obtained by CAROT as training targets, internalizing the aligned representations. The trained models outperform existing CLA methods in 11 of 18 evaluation settings (3 models \times 3 tasks \times ID/OOD languages). Our work provides insights into what constitutes effective alignment targets for CLA in LLMs. Code is available at https://github.com/ynklab/CAROT.</abstract>
    <identifier type="citekey">yamamoto-etal-2026-cross</identifier>
    <location>
        <url>https://aclanthology.org/2026.emnlp-main.933/</url>
    </location>
    <part>
        <date>2026-10</date>
        <extent unit="page">
            <start>21495</start>
            <end>21514</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Cross-Lingual Representation Alignment by Token-Level Optimal Transport in a Language-Agnostic Space
%A Yamamoto, Taisei
%A Kumon, Ryoma
%A Bollegala, Danushka
%A Yanaka, Hitomi
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-459-0
%F yamamoto-etal-2026-cross
%X Cross-lingual alignment (CLA) aims to align the representations of large language models (LLMs) across languages, enabling cross-lingual transfer to improve multilingual capabilities. Previous CLA methods often ignore language-specific information encoded in representations and only consider sentence-level alignment, which may lead to suboptimal performance and input-output language mismatch. We propose **CAROT** (**C**ross-Lingual **A**lignment of **R**epresentations in a Language-Agnostic Space via **O**ptimal **T**ransport), which consists of two steps: identifying language-specific representations in LLMs’ internal states and aligning language-agnostic representations across languages at the token level by optimal transport, while explicitly preserving language-specific representations. Inference-time steering experiments show that the representations computed by CAROT are effective alignment targets, improving multilingual performance by up to 11.2 points in accuracy while maintaining input-output language consistency. We further use the representations obtained by CAROT as training targets, internalizing the aligned representations. The trained models outperform existing CLA methods in 11 of 18 evaluation settings (3 models \times 3 tasks \times ID/OOD languages). Our work provides insights into what constitutes effective alignment targets for CLA in LLMs. Code is available at https://github.com/ynklab/CAROT.
%U https://aclanthology.org/2026.emnlp-main.933/
%P 21495-21514

Markdown (Informal)

[Cross-Lingual Representation Alignment by Token-Level Optimal Transport in a Language-Agnostic Space](https://aclanthology.org/2026.emnlp-main.933/) (Yamamoto et al., EMNLP 2026)

• Cross-Lingual Representation Alignment by Token-Level Optimal Transport in a Language-Agnostic Space (Yamamoto et al., EMNLP 2026)

ACL

• Taisei Yamamoto, Ryoma Kumon, Danushka Bollegala, and Hitomi Yanaka. 2026. Cross-Lingual Representation Alignment by Token-Level Optimal Transport in a Language-Agnostic Space. In Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing, pages 21495–21514, Budapest, Hungary. Association for Computational Linguistics.