@inproceedings{chua-etal-2026-rabakbench,
title = "{R}abak{B}ench: Evaluating Safety Guardrails in {S}ingapore{'}s Multilingual Vernaculars",
author = "Chua, Gabriel and
Tan, Leanne and
Ge, Ziyu and
Lee, Roy Ka-Wei",
editor = "Zhang, Yue and
Xu, Ruifeng and
Purwarianti, Ayu and
Rong, Tong and
Suzuki, Jun",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {IJCNLP}-{AACL} 2026",
month = nov,
year = "2026",
address = "Hengqin, China",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-aacl.38/",
pages = "729--753",
ISBN = "979-8-89176-444-6",
abstract = "LLM safety guardrails often behave unevenly in localized multilingual settings, especially with code-mixed vernaculars and non-Latin scripts. We introduce RabakBench, a Singapore-localized multilingual safety benchmark for Singlish, Chinese, Malay, and Tamil, built through a semi-automated, human-validated Generate{--}Label{--}Translate pipeline. The benchmark contains 1,341 unique examples in four language versions, yielding 5,364 entries across six harm categories, with LLM-assisted labels validated against human consensus. Evaluating 13 commercial and open-source guardrails reveals substantial cross-language and system-level variation, highlighting localized and script-sensitive safety gaps in the Singaporean context. RabakBench provides both a diagnostic resource and a reproducible, human-in-the-loop framework for constructing localized safety benchmarks."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="chua-etal-2026-rabakbench">
<titleInfo>
<title>RabakBench: Evaluating Safety Guardrails in Singapore’s Multilingual Vernaculars</title>
</titleInfo>
<name type="personal">
<namePart type="given">Gabriel</namePart>
<namePart type="family">Chua</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Leanne</namePart>
<namePart type="family">Tan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ziyu</namePart>
<namePart type="family">Ge</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Roy</namePart>
<namePart type="given">Ka-Wei</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-11</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: IJCNLP-AACL 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yue</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ruifeng</namePart>
<namePart type="family">Xu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ayu</namePart>
<namePart type="family">Purwarianti</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tong</namePart>
<namePart type="family">Rong</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jun</namePart>
<namePart type="family">Suzuki</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Hengqin, China</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-444-6</identifier>
</relatedItem>
<abstract>LLM safety guardrails often behave unevenly in localized multilingual settings, especially with code-mixed vernaculars and non-Latin scripts. We introduce RabakBench, a Singapore-localized multilingual safety benchmark for Singlish, Chinese, Malay, and Tamil, built through a semi-automated, human-validated Generate–Label–Translate pipeline. The benchmark contains 1,341 unique examples in four language versions, yielding 5,364 entries across six harm categories, with LLM-assisted labels validated against human consensus. Evaluating 13 commercial and open-source guardrails reveals substantial cross-language and system-level variation, highlighting localized and script-sensitive safety gaps in the Singaporean context. RabakBench provides both a diagnostic resource and a reproducible, human-in-the-loop framework for constructing localized safety benchmarks.</abstract>
<identifier type="citekey">chua-etal-2026-rabakbench</identifier>
<location>
<url>https://aclanthology.org/2026.findings-aacl.38/</url>
</location>
<part>
<date>2026-11</date>
<extent unit="page">
<start>729</start>
<end>753</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T RabakBench: Evaluating Safety Guardrails in Singapore’s Multilingual Vernaculars
%A Chua, Gabriel
%A Tan, Leanne
%A Ge, Ziyu
%A Lee, Roy Ka-Wei
%Y Zhang, Yue
%Y Xu, Ruifeng
%Y Purwarianti, Ayu
%Y Rong, Tong
%Y Suzuki, Jun
%S Findings of the Association for Computational Linguistics: IJCNLP-AACL 2026
%D 2026
%8 November
%I Association for Computational Linguistics
%C Hengqin, China
%@ 979-8-89176-444-6
%F chua-etal-2026-rabakbench
%X LLM safety guardrails often behave unevenly in localized multilingual settings, especially with code-mixed vernaculars and non-Latin scripts. We introduce RabakBench, a Singapore-localized multilingual safety benchmark for Singlish, Chinese, Malay, and Tamil, built through a semi-automated, human-validated Generate–Label–Translate pipeline. The benchmark contains 1,341 unique examples in four language versions, yielding 5,364 entries across six harm categories, with LLM-assisted labels validated against human consensus. Evaluating 13 commercial and open-source guardrails reveals substantial cross-language and system-level variation, highlighting localized and script-sensitive safety gaps in the Singaporean context. RabakBench provides both a diagnostic resource and a reproducible, human-in-the-loop framework for constructing localized safety benchmarks.
%U https://aclanthology.org/2026.findings-aacl.38/
%P 729-753
Markdown (Informal)
[RabakBench: Evaluating Safety Guardrails in Singapore’s Multilingual Vernaculars](https://aclanthology.org/2026.findings-aacl.38/) (Chua et al., Findings 2026)
ACL