@inproceedings{huang-etal-2026-msqa,
title = "{MSQA}: A Natively Sourced Multilingual and Multicultural {S}imple{QA} Benchmark",
author = "Huang, Yukai and
Chen, Xianru and
Chen, Mingxiang and
Lei, Xinping and
Deng, Fangbing and
Chen, Jin and
Liu, Jiaheng and
Zhang, Ge and
Huang, Wenhao",
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.1217/",
pages = "25777--25792",
ISBN = "979-8-89176-462-0",
abstract = "Multilingual fluency often invites a stronger assumption: a model that can speak a user{'}s language must also understand the culture encoded by that language. We call this the Illusion of Cultural Alignment. To test this assumption directly, we introduce MMC-QA, a benchmark of 1,086 natively sourced questions across 11 language groups, five cultural dimensions, and three difficulty tiers. Unlike translated benchmarks, MMC-QA targets locally grounded knowledge and reduces shortcuts from English-centric cross-lingual transfer. Evaluating 15 LLMs, we find substantial cultural degradation and a pronounced Locality Effect: cultural competence tracks pre-training exposure more closely than general reasoning ability. We further show that common inference-time remedies do not dissolve the illusion. Models remain overconfident on unfamiliar cultural questions, repeated sampling yields unstable rather than reliable correctness, and retrieval augmentation helps unevenly on long-tail facts. These findings indicate that cultural alignment cannot be inferred from multilingual ability alone and requires deeper intervention than calibration, sampling, or retrieval at inference time."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="huang-etal-2026-msqa">
<titleInfo>
<title>MSQA: A Natively Sourced Multilingual and Multicultural SimpleQA Benchmark</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yukai</namePart>
<namePart type="family">Huang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xianru</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mingxiang</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xinping</namePart>
<namePart type="family">Lei</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Fangbing</namePart>
<namePart type="family">Deng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jin</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jiaheng</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ge</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Wenhao</namePart>
<namePart type="family">Huang</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 fluency often invites a stronger assumption: a model that can speak a user’s language must also understand the culture encoded by that language. We call this the Illusion of Cultural Alignment. To test this assumption directly, we introduce MMC-QA, a benchmark of 1,086 natively sourced questions across 11 language groups, five cultural dimensions, and three difficulty tiers. Unlike translated benchmarks, MMC-QA targets locally grounded knowledge and reduces shortcuts from English-centric cross-lingual transfer. Evaluating 15 LLMs, we find substantial cultural degradation and a pronounced Locality Effect: cultural competence tracks pre-training exposure more closely than general reasoning ability. We further show that common inference-time remedies do not dissolve the illusion. Models remain overconfident on unfamiliar cultural questions, repeated sampling yields unstable rather than reliable correctness, and retrieval augmentation helps unevenly on long-tail facts. These findings indicate that cultural alignment cannot be inferred from multilingual ability alone and requires deeper intervention than calibration, sampling, or retrieval at inference time.</abstract>
<identifier type="citekey">huang-etal-2026-msqa</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.1217/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>25777</start>
<end>25792</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T MSQA: A Natively Sourced Multilingual and Multicultural SimpleQA Benchmark
%A Huang, Yukai
%A Chen, Xianru
%A Chen, Mingxiang
%A Lei, Xinping
%A Deng, Fangbing
%A Chen, Jin
%A Liu, Jiaheng
%A Zhang, Ge
%A Huang, Wenhao
%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 huang-etal-2026-msqa
%X Multilingual fluency often invites a stronger assumption: a model that can speak a user’s language must also understand the culture encoded by that language. We call this the Illusion of Cultural Alignment. To test this assumption directly, we introduce MMC-QA, a benchmark of 1,086 natively sourced questions across 11 language groups, five cultural dimensions, and three difficulty tiers. Unlike translated benchmarks, MMC-QA targets locally grounded knowledge and reduces shortcuts from English-centric cross-lingual transfer. Evaluating 15 LLMs, we find substantial cultural degradation and a pronounced Locality Effect: cultural competence tracks pre-training exposure more closely than general reasoning ability. We further show that common inference-time remedies do not dissolve the illusion. Models remain overconfident on unfamiliar cultural questions, repeated sampling yields unstable rather than reliable correctness, and retrieval augmentation helps unevenly on long-tail facts. These findings indicate that cultural alignment cannot be inferred from multilingual ability alone and requires deeper intervention than calibration, sampling, or retrieval at inference time.
%U https://aclanthology.org/2026.findings-emnlp.1217/
%P 25777-25792
Markdown (Informal)
[MSQA: A Natively Sourced Multilingual and Multicultural SimpleQA Benchmark](https://aclanthology.org/2026.findings-emnlp.1217/) (Huang et al., Findings 2026)
ACL
- Yukai Huang, Xianru Chen, Mingxiang Chen, Xinping Lei, Fangbing Deng, Jin Chen, Jiaheng Liu, Ge Zhang, and Wenhao Huang. 2026. MSQA: A Natively Sourced Multilingual and Multicultural SimpleQA Benchmark. In Findings of the Association for Computational Linguistics: EMNLP 2026, pages 25777–25792, Budapest, Hungary. Association for Computational Linguistics.