@inproceedings{rauniyar-etal-2026-pdos,
title = "{PDOS}: Widening the Safety Overton Window Through Persona-Driven Generation",
author = "Rauniyar, Kritesh and
Dras, Mark and
Naseem, Usman",
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.1077/",
pages = "22779--22803",
ISBN = "979-8-89176-462-0",
abstract = "Safety in large language models is often framed as requiring a single correct response, even in safety-sensitive contexts where multiple perspectives may exist. Existing approaches largely ignore Overton pluralism (inclusion of multiple reasonable viewpoints for a single query) and reduce complex context to monolithic judgments that fail to capture the range of reasonable viewpoints. Moreover, existing frameworks are not grounded in safety settings, which limits their ability to capture the diversity of perspectives required in safety-critical situations. Existing safety datasets do not provide Overton-style annotations, limiting their ability to represent multiple harm perspectives within the same safety-critical situation. To address this, we construct **SOS**, the safety-focused Overton dataset of 1,770 real-world safety situations annotated across 14 harm taxonomy. We further propose **PDOS**, a lightweight, training-free framework that retrieves personas from a synthetic pool and integrates their perspectives to generate unified Overton safety responses. PDOS achieves the highest average NLI coverage of 0.4226 among several baselines, and outperforms the strongest baseline by 13.3{\%}. Our findings show that Overton-based safety pluralism requires adaptable, context-aware approaches, providing insights into how these methods can better capture diverse perspectives."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="rauniyar-etal-2026-pdos">
<titleInfo>
<title>PDOS: Widening the Safety Overton Window Through Persona-Driven Generation</title>
</titleInfo>
<name type="personal">
<namePart type="given">Kritesh</namePart>
<namePart type="family">Rauniyar</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mark</namePart>
<namePart type="family">Dras</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Usman</namePart>
<namePart type="family">Naseem</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>Safety in large language models is often framed as requiring a single correct response, even in safety-sensitive contexts where multiple perspectives may exist. Existing approaches largely ignore Overton pluralism (inclusion of multiple reasonable viewpoints for a single query) and reduce complex context to monolithic judgments that fail to capture the range of reasonable viewpoints. Moreover, existing frameworks are not grounded in safety settings, which limits their ability to capture the diversity of perspectives required in safety-critical situations. Existing safety datasets do not provide Overton-style annotations, limiting their ability to represent multiple harm perspectives within the same safety-critical situation. To address this, we construct **SOS**, the safety-focused Overton dataset of 1,770 real-world safety situations annotated across 14 harm taxonomy. We further propose **PDOS**, a lightweight, training-free framework that retrieves personas from a synthetic pool and integrates their perspectives to generate unified Overton safety responses. PDOS achieves the highest average NLI coverage of 0.4226 among several baselines, and outperforms the strongest baseline by 13.3%. Our findings show that Overton-based safety pluralism requires adaptable, context-aware approaches, providing insights into how these methods can better capture diverse perspectives.</abstract>
<identifier type="citekey">rauniyar-etal-2026-pdos</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.1077/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>22779</start>
<end>22803</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T PDOS: Widening the Safety Overton Window Through Persona-Driven Generation
%A Rauniyar, Kritesh
%A Dras, Mark
%A Naseem, Usman
%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 rauniyar-etal-2026-pdos
%X Safety in large language models is often framed as requiring a single correct response, even in safety-sensitive contexts where multiple perspectives may exist. Existing approaches largely ignore Overton pluralism (inclusion of multiple reasonable viewpoints for a single query) and reduce complex context to monolithic judgments that fail to capture the range of reasonable viewpoints. Moreover, existing frameworks are not grounded in safety settings, which limits their ability to capture the diversity of perspectives required in safety-critical situations. Existing safety datasets do not provide Overton-style annotations, limiting their ability to represent multiple harm perspectives within the same safety-critical situation. To address this, we construct **SOS**, the safety-focused Overton dataset of 1,770 real-world safety situations annotated across 14 harm taxonomy. We further propose **PDOS**, a lightweight, training-free framework that retrieves personas from a synthetic pool and integrates their perspectives to generate unified Overton safety responses. PDOS achieves the highest average NLI coverage of 0.4226 among several baselines, and outperforms the strongest baseline by 13.3%. Our findings show that Overton-based safety pluralism requires adaptable, context-aware approaches, providing insights into how these methods can better capture diverse perspectives.
%U https://aclanthology.org/2026.findings-emnlp.1077/
%P 22779-22803
Markdown (Informal)
[PDOS: Widening the Safety Overton Window Through Persona-Driven Generation](https://aclanthology.org/2026.findings-emnlp.1077/) (Rauniyar et al., Findings 2026)
ACL