@inproceedings{mitra-agrawal-2026-choosing,
title = "Choosing the Right Language Mode at Inference Time for Multilingual Reliability",
author = "Mitra, Ekata and
Agrawal, Ameeta",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {EMNLP} 2026",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-emnlp.89/",
pages = "1821--1839",
ISBN = "979-8-89176-462-0",
abstract = "Multilingual large language models often struggle to reason in low- to mid-resource languages. Prior work has shown that translation can improve multilingual reasoning by helping models access stronger English-centric representations. This raises a central question: how much translation is needed for multilingual large language models to reason reliably, and when does additional translation instead trigger interference and overconfidence? Using LLaMA and Qwen models, we conduct extensive experiments that vary both text scope and language mode (target-only, English-only, and bilingual) to evaluate accuracy and reliability. Our results reveal a clear trade-off: English context often improves understanding and helps recover errors caused by non-English comprehension, yet adding redundant bilingual context can intensify interference. To address this trade-off, we propose Reliability-Aware Adaptive Inference (RAAI), a training-free test-time framework that (i) performs ECE-aware routing and prompt fusion and (ii) uses a mid-layer Risk Index (RI) to gate sequential reasoning, allocating extra compute only when it is likely to help while suppressing harmful bilingual redundancy. Across two model families, RAAI improves accuracy by 25 to 37.7 points in low-resource languages and reduces calibration error by about 3 to 6 points, with the largest gains appearing in the lowest-resource language tiers."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="mitra-agrawal-2026-choosing">
<titleInfo>
<title>Choosing the Right Language Mode at Inference Time for Multilingual Reliability</title>
</titleInfo>
<name type="personal">
<namePart type="given">Ekata</namePart>
<namePart type="family">Mitra</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ameeta</namePart>
<namePart type="family">Agrawal</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>Findings of the Association for Computational Linguistics: EMNLP 2026</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">979-8-89176-462-0</identifier>
</relatedItem>
<abstract>Multilingual large language models often struggle to reason in low- to mid-resource languages. Prior work has shown that translation can improve multilingual reasoning by helping models access stronger English-centric representations. This raises a central question: how much translation is needed for multilingual large language models to reason reliably, and when does additional translation instead trigger interference and overconfidence? Using LLaMA and Qwen models, we conduct extensive experiments that vary both text scope and language mode (target-only, English-only, and bilingual) to evaluate accuracy and reliability. Our results reveal a clear trade-off: English context often improves understanding and helps recover errors caused by non-English comprehension, yet adding redundant bilingual context can intensify interference. To address this trade-off, we propose Reliability-Aware Adaptive Inference (RAAI), a training-free test-time framework that (i) performs ECE-aware routing and prompt fusion and (ii) uses a mid-layer Risk Index (RI) to gate sequential reasoning, allocating extra compute only when it is likely to help while suppressing harmful bilingual redundancy. Across two model families, RAAI improves accuracy by 25 to 37.7 points in low-resource languages and reduces calibration error by about 3 to 6 points, with the largest gains appearing in the lowest-resource language tiers.</abstract>
<identifier type="citekey">mitra-agrawal-2026-choosing</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.89/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>1821</start>
<end>1839</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Choosing the Right Language Mode at Inference Time for Multilingual Reliability
%A Mitra, Ekata
%A Agrawal, Ameeta
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Findings of the Association for Computational Linguistics: EMNLP 2026
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-462-0
%F mitra-agrawal-2026-choosing
%X Multilingual large language models often struggle to reason in low- to mid-resource languages. Prior work has shown that translation can improve multilingual reasoning by helping models access stronger English-centric representations. This raises a central question: how much translation is needed for multilingual large language models to reason reliably, and when does additional translation instead trigger interference and overconfidence? Using LLaMA and Qwen models, we conduct extensive experiments that vary both text scope and language mode (target-only, English-only, and bilingual) to evaluate accuracy and reliability. Our results reveal a clear trade-off: English context often improves understanding and helps recover errors caused by non-English comprehension, yet adding redundant bilingual context can intensify interference. To address this trade-off, we propose Reliability-Aware Adaptive Inference (RAAI), a training-free test-time framework that (i) performs ECE-aware routing and prompt fusion and (ii) uses a mid-layer Risk Index (RI) to gate sequential reasoning, allocating extra compute only when it is likely to help while suppressing harmful bilingual redundancy. Across two model families, RAAI improves accuracy by 25 to 37.7 points in low-resource languages and reduces calibration error by about 3 to 6 points, with the largest gains appearing in the lowest-resource language tiers.
%U https://aclanthology.org/2026.findings-emnlp.89/
%P 1821-1839
Markdown (Informal)
[Choosing the Right Language Mode at Inference Time for Multilingual Reliability](https://aclanthology.org/2026.findings-emnlp.89/) (Mitra & Agrawal, Findings 2026)
ACL