@inproceedings{mundada-etal-2026-evaluating,
title = "Evaluating Language Model Pluralism through In-the-wild Crowd Discussions",
author = "Mundada, Gagan and
Surana, Rohan and
Swaminathan, Nandhini and
Majumder, Bodhisattwa Prasad and
Wu, Junda and
McAuley, Julian and
Xie, Zhouhang",
editor = "Liakata, Maria and
Moreira, Viviane P. and
Zhang, Jiajun and
Jurgens, David",
booktitle = "Proceedings of the 64th Annual Meeting of the {A}ssociation for {C}omputational {L}inguistics (Volume 1: Long Papers)",
month = jul,
year = "2026",
address = "San Diego, California, United States",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.acl-long.1957/",
pages = "42273--42296",
ISBN = "979-8-89176-390-6",
abstract = "When answering subjective questions, an ideal LLM should surface diverse plausible perspectives rather than favoring a single viewpoint, a characteristic known as pluralism. Recent studies show that modern LLMs optimized through preference alignment systematically favor certain positions on subjective queries, making pluralism evaluation increasingly important. However, existing evaluation methods focus dominantly on multiple-choice and question-answering tasks, leaving open-ended generation largely unaddressed.We propose PLURALEVAL, an evaluation framework that assesses LLM pluralism in open-ended generation by comparing outputs against free-form crowd responses. Our approach decomposes ground-truth responses into atomic, non-overlapping claims, then evaluates whether LLMs adequately cover this diverse claim space. We then introduce WildSCOPE, a multi-domain dataset of natural crowd responses, and demonstrate that PLURALEVAL captures novel insights, such as the collapse of pluralism through sycophancy, where LLM systematically degrades in overton pluralism when a user{'}s belief is revealed. Finally, we discuss the value and actionable insights for preserving and encouraging pluralism from LLM deployers' side."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="mundada-etal-2026-evaluating">
<titleInfo>
<title>Evaluating Language Model Pluralism through In-the-wild Crowd Discussions</title>
</titleInfo>
<name type="personal">
<namePart type="given">Gagan</namePart>
<namePart type="family">Mundada</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Rohan</namePart>
<namePart type="family">Surana</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Nandhini</namePart>
<namePart type="family">Swaminathan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Bodhisattwa</namePart>
<namePart type="given">Prasad</namePart>
<namePart type="family">Majumder</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Junda</namePart>
<namePart type="family">Wu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Julian</namePart>
<namePart type="family">McAuley</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhouhang</namePart>
<namePart type="family">Xie</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Maria</namePart>
<namePart type="family">Liakata</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Viviane</namePart>
<namePart type="given">P</namePart>
<namePart type="family">Moreira</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jiajun</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">David</namePart>
<namePart type="family">Jurgens</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">San Diego, California, United States</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-390-6</identifier>
</relatedItem>
<abstract>When answering subjective questions, an ideal LLM should surface diverse plausible perspectives rather than favoring a single viewpoint, a characteristic known as pluralism. Recent studies show that modern LLMs optimized through preference alignment systematically favor certain positions on subjective queries, making pluralism evaluation increasingly important. However, existing evaluation methods focus dominantly on multiple-choice and question-answering tasks, leaving open-ended generation largely unaddressed.We propose PLURALEVAL, an evaluation framework that assesses LLM pluralism in open-ended generation by comparing outputs against free-form crowd responses. Our approach decomposes ground-truth responses into atomic, non-overlapping claims, then evaluates whether LLMs adequately cover this diverse claim space. We then introduce WildSCOPE, a multi-domain dataset of natural crowd responses, and demonstrate that PLURALEVAL captures novel insights, such as the collapse of pluralism through sycophancy, where LLM systematically degrades in overton pluralism when a user’s belief is revealed. Finally, we discuss the value and actionable insights for preserving and encouraging pluralism from LLM deployers’ side.</abstract>
<identifier type="citekey">mundada-etal-2026-evaluating</identifier>
<location>
<url>https://aclanthology.org/2026.acl-long.1957/</url>
</location>
<part>
<date>2026-07</date>
<extent unit="page">
<start>42273</start>
<end>42296</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Evaluating Language Model Pluralism through In-the-wild Crowd Discussions
%A Mundada, Gagan
%A Surana, Rohan
%A Swaminathan, Nandhini
%A Majumder, Bodhisattwa Prasad
%A Wu, Junda
%A McAuley, Julian
%A Xie, Zhouhang
%Y Liakata, Maria
%Y Moreira, Viviane P.
%Y Zhang, Jiajun
%Y Jurgens, David
%S Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
%D 2026
%8 July
%I Association for Computational Linguistics
%C San Diego, California, United States
%@ 979-8-89176-390-6
%F mundada-etal-2026-evaluating
%X When answering subjective questions, an ideal LLM should surface diverse plausible perspectives rather than favoring a single viewpoint, a characteristic known as pluralism. Recent studies show that modern LLMs optimized through preference alignment systematically favor certain positions on subjective queries, making pluralism evaluation increasingly important. However, existing evaluation methods focus dominantly on multiple-choice and question-answering tasks, leaving open-ended generation largely unaddressed.We propose PLURALEVAL, an evaluation framework that assesses LLM pluralism in open-ended generation by comparing outputs against free-form crowd responses. Our approach decomposes ground-truth responses into atomic, non-overlapping claims, then evaluates whether LLMs adequately cover this diverse claim space. We then introduce WildSCOPE, a multi-domain dataset of natural crowd responses, and demonstrate that PLURALEVAL captures novel insights, such as the collapse of pluralism through sycophancy, where LLM systematically degrades in overton pluralism when a user’s belief is revealed. Finally, we discuss the value and actionable insights for preserving and encouraging pluralism from LLM deployers’ side.
%U https://aclanthology.org/2026.acl-long.1957/
%P 42273-42296
Markdown (Informal)
[Evaluating Language Model Pluralism through In-the-wild Crowd Discussions](https://aclanthology.org/2026.acl-long.1957/) (Mundada et al., ACL 2026)
ACL
- Gagan Mundada, Rohan Surana, Nandhini Swaminathan, Bodhisattwa Prasad Majumder, Junda Wu, Julian McAuley, and Zhouhang Xie. 2026. Evaluating Language Model Pluralism through In-the-wild Crowd Discussions. In Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 42273–42296, San Diego, California, United States. Association for Computational Linguistics.