@inproceedings{garg-etal-2026-stable,
title = "Stable Scores, Failed Retrieval: Diagnosing {RAG} Evaluation in Low-Resource {I}ndic Languages",
author = "Garg, Priyanshi and
Gustafson, Jasmin and
Fan, Yingzhe",
editor = "Wang, Yimu and
Choi, Yee Man and
Wu, Di and
Ouyang, Siqi and
Azadani, Mozhgan Nasr and
Fung, Yi R. (May)",
booktitle = "Proceedings of the 1st Workshop on Grounding Language Models: Learning Faithfully and Efficiently ({G}round{LM} 2026)",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.groundlm-1.48/",
pages = "637--648",
ISBN = "979-8-89176-478-1",
abstract = "Automatic metrics stand in for native-speaker evaluation in low-resource retrieval-augmented generation (RAG), so their validity determines what can be concluded about these systems. We test whether common metrics detect re- trieval and grounding failure in one fixed RAG pipeline across ten Indic languages. They do not merely miss the failure in Santali; they invert it. Hit@5 is 0.011, yet raw XLM-R BERTScore is the highest in the set at 0.902, changes little when retrieved evidence is re- moved, and assigns nearly the same score to matched and unrelated references. The cause is coverage: the scorer maps 39{\%} of Santali tokens to {\ensuremath{<}}unk{\ensuremath{>}}, and a scorer whose vocab- ulary includes Ol Chiki restores discrimina- tion while LaBSE does not. Retrieval fails for the same reason. The gold article is indexed for 80{\%} of Santali queries, but the retriever shares XLM-R{'}s tokenizer and collapses before reranking. Judge-based faithfulness and con- text precision, computed by a model with a dif- ferent tokenizer, flag the failure independently. A metric that cannot represent a script does not become noisy; it becomes constant and high. We argue that metric validity in low-resource languages must be established per language, and that evidence removal, reference shuffling, and the scorer{'}s {\ensuremath{<}}unk{\ensuremath{>}} rate can be important additions."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="garg-etal-2026-stable">
<titleInfo>
<title>Stable Scores, Failed Retrieval: Diagnosing RAG Evaluation in Low-Resource Indic Languages</title>
</titleInfo>
<name type="personal">
<namePart type="given">Priyanshi</namePart>
<namePart type="family">Garg</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jasmin</namePart>
<namePart type="family">Gustafson</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yingzhe</namePart>
<namePart type="family">Fan</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 1st Workshop on Grounding Language Models: Learning Faithfully and Efficiently (GroundLM 2026)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yimu</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yee</namePart>
<namePart type="given">Man</namePart>
<namePart type="family">Choi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Di</namePart>
<namePart type="family">Wu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Siqi</namePart>
<namePart type="family">Ouyang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mozhgan</namePart>
<namePart type="given">Nasr</namePart>
<namePart type="family">Azadani</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yi</namePart>
<namePart type="given">R</namePart>
<namePart type="given">(May)</namePart>
<namePart type="family">Fung</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">979-8-89176-478-1</identifier>
</relatedItem>
<abstract>Automatic metrics stand in for native-speaker evaluation in low-resource retrieval-augmented generation (RAG), so their validity determines what can be concluded about these systems. We test whether common metrics detect re- trieval and grounding failure in one fixed RAG pipeline across ten Indic languages. They do not merely miss the failure in Santali; they invert it. Hit@5 is 0.011, yet raw XLM-R BERTScore is the highest in the set at 0.902, changes little when retrieved evidence is re- moved, and assigns nearly the same score to matched and unrelated references. The cause is coverage: the scorer maps 39% of Santali tokens to \ensuremath<unk\ensuremath>, and a scorer whose vocab- ulary includes Ol Chiki restores discrimina- tion while LaBSE does not. Retrieval fails for the same reason. The gold article is indexed for 80% of Santali queries, but the retriever shares XLM-R’s tokenizer and collapses before reranking. Judge-based faithfulness and con- text precision, computed by a model with a dif- ferent tokenizer, flag the failure independently. A metric that cannot represent a script does not become noisy; it becomes constant and high. We argue that metric validity in low-resource languages must be established per language, and that evidence removal, reference shuffling, and the scorer’s \ensuremath<unk\ensuremath> rate can be important additions.</abstract>
<identifier type="citekey">garg-etal-2026-stable</identifier>
<location>
<url>https://aclanthology.org/2026.groundlm-1.48/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>637</start>
<end>648</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Stable Scores, Failed Retrieval: Diagnosing RAG Evaluation in Low-Resource Indic Languages
%A Garg, Priyanshi
%A Gustafson, Jasmin
%A Fan, Yingzhe
%Y Wang, Yimu
%Y Choi, Yee Man
%Y Wu, Di
%Y Ouyang, Siqi
%Y Azadani, Mozhgan Nasr
%Y Fung, Yi R. (May)
%S Proceedings of the 1st Workshop on Grounding Language Models: Learning Faithfully and Efficiently (GroundLM 2026)
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-478-1
%F garg-etal-2026-stable
%X Automatic metrics stand in for native-speaker evaluation in low-resource retrieval-augmented generation (RAG), so their validity determines what can be concluded about these systems. We test whether common metrics detect re- trieval and grounding failure in one fixed RAG pipeline across ten Indic languages. They do not merely miss the failure in Santali; they invert it. Hit@5 is 0.011, yet raw XLM-R BERTScore is the highest in the set at 0.902, changes little when retrieved evidence is re- moved, and assigns nearly the same score to matched and unrelated references. The cause is coverage: the scorer maps 39% of Santali tokens to \ensuremath<unk\ensuremath>, and a scorer whose vocab- ulary includes Ol Chiki restores discrimina- tion while LaBSE does not. Retrieval fails for the same reason. The gold article is indexed for 80% of Santali queries, but the retriever shares XLM-R’s tokenizer and collapses before reranking. Judge-based faithfulness and con- text precision, computed by a model with a dif- ferent tokenizer, flag the failure independently. A metric that cannot represent a script does not become noisy; it becomes constant and high. We argue that metric validity in low-resource languages must be established per language, and that evidence removal, reference shuffling, and the scorer’s \ensuremath<unk\ensuremath> rate can be important additions.
%U https://aclanthology.org/2026.groundlm-1.48/
%P 637-648
Markdown (Informal)
[Stable Scores, Failed Retrieval: Diagnosing RAG Evaluation in Low-Resource Indic Languages](https://aclanthology.org/2026.groundlm-1.48/) (Garg et al., GroundLM 2026)
ACL