<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="prama-etal-2026-translation">
    <titleInfo>
        <title>Translation-CoT: A Human-Inspired Chain-of-Thought Framework for Multilingual LLM Translation</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Tabia</namePart>
        <namePart type="given">Tanzin</namePart>
        <namePart type="family">Prama</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Juniper</namePart>
        <namePart type="given">L</namePart>
        <namePart type="family">Lovato</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Chris</namePart>
        <namePart type="family">Danforth</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Peter</namePart>
        <namePart type="family">Dodds</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2026-08</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Proceedings of the 17th Conference of the Association for Machine Translation in the Americas (Volume 1: Research Track)</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Eleftheria</namePart>
            <namePart type="family">Briakou</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Jeremy</namePart>
            <namePart type="family">Gwinnup</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Shivali</namePart>
            <namePart type="family">Goel</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <originInfo>
            <publisher>Association for Machine Translation in the Americas</publisher>
            <place>
                <placeTerm type="text">Québec City, Canada</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
    </relatedItem>
    <abstract>Large language models (LLMs) have transformed machine translation, yet mistranslations, hallucinations, and unnatural phrasing still limit their effectiveness, particularly for low-resource languages. We propose Translation-CoT, a chain-of-thought prompting strategy that breaks translation into structured stages (lexical retrieval, grammatical analysis, and topic identification), followed by a refinement step to improve fluency, tone, and idiomatic expression. We evaluate Translation-CoT across 14 languages from 14 language families and multiple LLMs (GPT-4o, GPT-4o-mini, LLaMA 3.1, and Gemma 2), with GPT-4o performing best overall, in both English łeftrightarrow non-English (X) translation settings. Compared with zero-shot prompting, in-context learning, and existing chain-of-thought prompting methods (Tree-of-Thought (ToT) and Learning-Oriented Prompting (LOT)), Translation-CoT outperforms these prompting strategies on multilingual machine translation across BLEU, ChrF, and METEOR, with especially strong gains in the more difficult English→non-English (X) setting and in low-resource languages. Human evaluation further shows higher preference scores and lower MQM penalty scores, indicating fewer mistranslations, omissions, awkward phrasing, and hallucinations with Translation-CoT. Overall, our results show that structured, task-aware prompting is an effective approach for improving multilingual translation quality and robustness in LLMs.</abstract>
    <identifier type="citekey">prama-etal-2026-translation</identifier>
    <location>
        <url>https://aclanthology.org/2026.amta-research.2/</url>
    </location>
    <part>
        <date>2026-08</date>
        <extent unit="page">
            <start>3</start>
            <end>27</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Translation-CoT: A Human-Inspired Chain-of-Thought Framework for Multilingual LLM Translation
%A Prama, Tabia Tanzin
%A Lovato, Juniper L.
%A Danforth, Chris
%A Dodds, Peter
%Y Briakou, Eleftheria
%Y Gwinnup, Jeremy
%Y Goel, Shivali
%S Proceedings of the 17th Conference of the Association for Machine Translation in the Americas (Volume 1: Research Track)
%D 2026
%8 August
%I Association for Machine Translation in the Americas
%C Québec City, Canada
%F prama-etal-2026-translation
%X Large language models (LLMs) have transformed machine translation, yet mistranslations, hallucinations, and unnatural phrasing still limit their effectiveness, particularly for low-resource languages. We propose Translation-CoT, a chain-of-thought prompting strategy that breaks translation into structured stages (lexical retrieval, grammatical analysis, and topic identification), followed by a refinement step to improve fluency, tone, and idiomatic expression. We evaluate Translation-CoT across 14 languages from 14 language families and multiple LLMs (GPT-4o, GPT-4o-mini, LLaMA 3.1, and Gemma 2), with GPT-4o performing best overall, in both English łeftrightarrow non-English (X) translation settings. Compared with zero-shot prompting, in-context learning, and existing chain-of-thought prompting methods (Tree-of-Thought (ToT) and Learning-Oriented Prompting (LOT)), Translation-CoT outperforms these prompting strategies on multilingual machine translation across BLEU, ChrF, and METEOR, with especially strong gains in the more difficult English→non-English (X) setting and in low-resource languages. Human evaluation further shows higher preference scores and lower MQM penalty scores, indicating fewer mistranslations, omissions, awkward phrasing, and hallucinations with Translation-CoT. Overall, our results show that structured, task-aware prompting is an effective approach for improving multilingual translation quality and robustness in LLMs.
%U https://aclanthology.org/2026.amta-research.2/
%P 3-27

Markdown (Informal)

[Translation-CoT: A Human-Inspired Chain-of-Thought Framework for Multilingual LLM Translation](https://aclanthology.org/2026.amta-research.2/) (Prama et al., AMTA 2026)

• Translation-CoT: A Human-Inspired Chain-of-Thought Framework for Multilingual LLM Translation (Prama et al., AMTA 2026)

ACL

• Tabia Tanzin Prama, Juniper L Lovato, Chris Danforth, and Peter Dodds. 2026. Translation-CoT: A Human-Inspired Chain-of-Thought Framework for Multilingual LLM Translation. In Proceedings of the 17th Conference of the Association for Machine Translation in the Americas (Volume 1: Research Track), pages 3–27, Québec City, Canada. Association for Machine Translation in the Americas.