@inproceedings{ozturk-etal-2026-nedoqwen,
title = "{NEDOQ}wen: Diagnosing and Repairing a {T}urkish-Centric Small Language Model for Multilingual Evaluation",
author = {{\"O}zt{\"u}rk, Ahmet R{\i}fat and
Dalar, Ya{\u{g}}{\i}z Ekrem and
Sezer, Nedim Mutlu and
Aksoy, {\"O}mer Faruk and
Saliho{\u{g}}lu, Feyzi Arda},
editor = "Adelani, David Ifeoluwa and
Alabi, Jesujoba and
Arnett, Catherine and
Chang, Tyler and
Downey, C.M. and
Goldman, Omer and
Li, Senyu and
Miranda, Lester James Validad and
Qi, Jirui and
Stap, David and
Winata, Genta Indra and
Zhang, Xinyu Crystina",
booktitle = "Proceedings of the 6th Workshop on Multilingual Representation Learning ({MRL} 2026)",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.mrl-1.2/",
pages = "19--26",
ISBN = "979-8-89176-469-9",
abstract = "Multilingual language modeling is often evaluated through broad aggregate scores, yet language-specific failures can remain hidden, especially for morphologically rich and underrepresented languages. We present NEDOQwen, a Turkish-centric 0.824B decoder-only small language model research artifact, and use it to study a low-cost diagnosis-and-repair workflow for Turkish evaluation. Internal Turkish instruction diagnostics improve after supervised fine-tuning, but external Turkish and Turkic multiple-choice benchmarks reveal weak academic knowledge and severe answer-label bias. A compact benchmark-safe repair stage yields a small improvement on TurkishMMLU-sub, from 18.11{\%} to 21.00{\%}, and a larger improvement on TUMLU-mini Turkish, from 21.33{\%} to 32.22{\%}, while an answer-label-only ablation reaches 19.56{\%} and 27.00{\%}. We further analyze tokenizer coverage, internal-vs-external evaluation mismatch, prediction-label distributions, paired uncertainty, and compute cost. The contribution is not a state-of-the-art Turkish model, but an auditable case study showing how small language-specific models can expose and partially repair multilingual evaluation failures while remaining unreliable."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="ozturk-etal-2026-nedoqwen">
<titleInfo>
<title>NEDOQwen: Diagnosing and Repairing a Turkish-Centric Small Language Model for Multilingual Evaluation</title>
</titleInfo>
<name type="personal">
<namePart type="given">Ahmet</namePart>
<namePart type="given">Rıfat</namePart>
<namePart type="family">Öztürk</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yağız</namePart>
<namePart type="given">Ekrem</namePart>
<namePart type="family">Dalar</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Nedim</namePart>
<namePart type="given">Mutlu</namePart>
<namePart type="family">Sezer</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ömer</namePart>
<namePart type="given">Faruk</namePart>
<namePart type="family">Aksoy</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Feyzi</namePart>
<namePart type="given">Arda</namePart>
<namePart type="family">Salihoğlu</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 6th Workshop on Multilingual Representation Learning (MRL 2026)</title>
</titleInfo>
<name type="personal">
<namePart type="given">David</namePart>
<namePart type="given">Ifeoluwa</namePart>
<namePart type="family">Adelani</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jesujoba</namePart>
<namePart type="family">Alabi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Catherine</namePart>
<namePart type="family">Arnett</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tyler</namePart>
<namePart type="family">Chang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">C</namePart>
<namePart type="given">M</namePart>
<namePart type="family">Downey</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Omer</namePart>
<namePart type="family">Goldman</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Senyu</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Lester</namePart>
<namePart type="given">James</namePart>
<namePart type="given">Validad</namePart>
<namePart type="family">Miranda</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jirui</namePart>
<namePart type="family">Qi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">David</namePart>
<namePart type="family">Stap</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Genta</namePart>
<namePart type="given">Indra</namePart>
<namePart type="family">Winata</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xinyu</namePart>
<namePart type="given">Crystina</namePart>
<namePart type="family">Zhang</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">979-8-89176-469-9</identifier>
</relatedItem>
<abstract>Multilingual language modeling is often evaluated through broad aggregate scores, yet language-specific failures can remain hidden, especially for morphologically rich and underrepresented languages. We present NEDOQwen, a Turkish-centric 0.824B decoder-only small language model research artifact, and use it to study a low-cost diagnosis-and-repair workflow for Turkish evaluation. Internal Turkish instruction diagnostics improve after supervised fine-tuning, but external Turkish and Turkic multiple-choice benchmarks reveal weak academic knowledge and severe answer-label bias. A compact benchmark-safe repair stage yields a small improvement on TurkishMMLU-sub, from 18.11% to 21.00%, and a larger improvement on TUMLU-mini Turkish, from 21.33% to 32.22%, while an answer-label-only ablation reaches 19.56% and 27.00%. We further analyze tokenizer coverage, internal-vs-external evaluation mismatch, prediction-label distributions, paired uncertainty, and compute cost. The contribution is not a state-of-the-art Turkish model, but an auditable case study showing how small language-specific models can expose and partially repair multilingual evaluation failures while remaining unreliable.</abstract>
<identifier type="citekey">ozturk-etal-2026-nedoqwen</identifier>
<location>
<url>https://aclanthology.org/2026.mrl-1.2/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>19</start>
<end>26</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T NEDOQwen: Diagnosing and Repairing a Turkish-Centric Small Language Model for Multilingual Evaluation
%A Öztürk, Ahmet Rıfat
%A Dalar, Yağız Ekrem
%A Sezer, Nedim Mutlu
%A Aksoy, Ömer Faruk
%A Salihoğlu, Feyzi Arda
%Y Adelani, David Ifeoluwa
%Y Alabi, Jesujoba
%Y Arnett, Catherine
%Y Chang, Tyler
%Y Downey, C. M.
%Y Goldman, Omer
%Y Li, Senyu
%Y Miranda, Lester James Validad
%Y Qi, Jirui
%Y Stap, David
%Y Winata, Genta Indra
%Y Zhang, Xinyu Crystina
%S Proceedings of the 6th Workshop on Multilingual Representation Learning (MRL 2026)
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-469-9
%F ozturk-etal-2026-nedoqwen
%X Multilingual language modeling is often evaluated through broad aggregate scores, yet language-specific failures can remain hidden, especially for morphologically rich and underrepresented languages. We present NEDOQwen, a Turkish-centric 0.824B decoder-only small language model research artifact, and use it to study a low-cost diagnosis-and-repair workflow for Turkish evaluation. Internal Turkish instruction diagnostics improve after supervised fine-tuning, but external Turkish and Turkic multiple-choice benchmarks reveal weak academic knowledge and severe answer-label bias. A compact benchmark-safe repair stage yields a small improvement on TurkishMMLU-sub, from 18.11% to 21.00%, and a larger improvement on TUMLU-mini Turkish, from 21.33% to 32.22%, while an answer-label-only ablation reaches 19.56% and 27.00%. We further analyze tokenizer coverage, internal-vs-external evaluation mismatch, prediction-label distributions, paired uncertainty, and compute cost. The contribution is not a state-of-the-art Turkish model, but an auditable case study showing how small language-specific models can expose and partially repair multilingual evaluation failures while remaining unreliable.
%U https://aclanthology.org/2026.mrl-1.2/
%P 19-26
Markdown (Informal)
[NEDOQwen: Diagnosing and Repairing a Turkish-Centric Small Language Model for Multilingual Evaluation](https://aclanthology.org/2026.mrl-1.2/) (Öztürk et al., MRL 2026)
ACL