<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="attri-attri-2026-intent">
    <titleInfo>
        <title>Same Intent, Different Script: Cross-Script Safety Degradation in LLMs Across Punjabi Sociolinguistic Variants</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Arnav</namePart>
        <namePart type="family">Attri</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Anuj</namePart>
        <namePart type="family">Attri</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 safety alignment typically assumes standard, canonical scripts, overlooking orthographic variation. We introduce PunjabiSafetyBench, a parallel safety benchmark evaluating 300 harmful intents and 500 benign controls across six sociolinguistic script variants (Gurmukhi, Shahmukhi, Romanized, code-mixed, Urdu, English). Evaluating six frontier LLMs, we reveal severe safety degradation on non-canonical scripts: transitioning queries to Romanized Punjabi or Shahmukhi drives significant guardrail bypasses in open-weight models, while proprietary models remain robust. Furthermore, we expose a localization gap where models overlook regional harmful slang despite successfully refusing standard clinical equivalents. Our findings demonstrate that modern guardrails are fragile to surface orthography, necessitating safety alignment on informal, non-standard scripts.</abstract>
    <identifier type="citekey">attri-attri-2026-intent</identifier>
    <location>
        <url>https://aclanthology.org/2026.findings-emnlp.362/</url>
    </location>
    <part>
        <date>2026-10</date>
        <extent unit="page">
            <start>7575</start>
            <end>7593</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Same Intent, Different Script: Cross-Script Safety Degradation in LLMs Across Punjabi Sociolinguistic Variants
%A Attri, Arnav
%A Attri, Anuj
%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 attri-attri-2026-intent
%X Multilingual safety alignment typically assumes standard, canonical scripts, overlooking orthographic variation. We introduce PunjabiSafetyBench, a parallel safety benchmark evaluating 300 harmful intents and 500 benign controls across six sociolinguistic script variants (Gurmukhi, Shahmukhi, Romanized, code-mixed, Urdu, English). Evaluating six frontier LLMs, we reveal severe safety degradation on non-canonical scripts: transitioning queries to Romanized Punjabi or Shahmukhi drives significant guardrail bypasses in open-weight models, while proprietary models remain robust. Furthermore, we expose a localization gap where models overlook regional harmful slang despite successfully refusing standard clinical equivalents. Our findings demonstrate that modern guardrails are fragile to surface orthography, necessitating safety alignment on informal, non-standard scripts.
%U https://aclanthology.org/2026.findings-emnlp.362/
%P 7575-7593

Markdown (Informal)

[Same Intent, Different Script: Cross-Script Safety Degradation in LLMs Across Punjabi Sociolinguistic Variants](https://aclanthology.org/2026.findings-emnlp.362/) (Attri & Attri, Findings 2026)

• Same Intent, Different Script: Cross-Script Safety Degradation in LLMs Across Punjabi Sociolinguistic Variants (Attri & Attri, Findings 2026)

ACL

• Arnav Attri and Anuj Attri. 2026. Same Intent, Different Script: Cross-Script Safety Degradation in LLMs Across Punjabi Sociolinguistic Variants. In Findings of the Association for Computational Linguistics: EMNLP 2026, pages 7575–7593, Budapest, Hungary. Association for Computational Linguistics.