@inproceedings{widanagamage-etal-2026-plurva,
title = "{P}lur{VA}-{LLM}-2026 Shared Task {T}rack-1: Curriculum Fine-Tuning with Soft Labels and Positional Consistency for Cross-Cultural Value Alignment",
author = "Widanagamage, Amasha and
Seneviratne, Hansika and
Fernando, Naween and
Pavithra, Dilushri and
Weerasinghe, Ruvan and
Sumanathilaka, Deshan",
editor = "Xiong, Deyi and
Branco, Ant{\'o}nio and
Zhang, Hongming and
Dong, Yue and
Wang, Benyou and
Zhang, Wenxuan and
Zhou, Li and
Zheng, Jingting",
booktitle = "Proceedings of the First Workshop on Pluralistic Value Alignment of {LLM}s",
month = nov,
year = "2026",
address = "Hengqin, China",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.plurvallm-1.14/",
pages = "187--196",
ISBN = "979-8-89176-453-8",
abstract = "Pluralistic value alignment asks whether a language model can adapt its judgements to distinct, locally grounded value systems rather than reflecting a single aggregated notion of human preference. The PlurVA-LLM shared task evaluates this setting by unifying three independently constructed, culturally grounded multiple-choice value dilemmas across Chinese, Indonesian, and Sri Lankan datasets. We present a single-model submission based on Qwen3.5-4B with LoRA fine-tuning, trained only on the 1,359 labelled development instances provided by the task. Our method combines three ideas motivated directly by the task structure: soft-label supervision for Indonesian annotator-vote distributions, consistency regularisation across option permutations to reduce answer-position bias, and a staged multilingual curriculum with cross-validation-based iteration selection. On the hidden test set, our system achieves a macro-average accuracy of 0.8062, with 0.7788 on Chinese, 0.7139 on Indonesian, and 0.9260 on Sri Lankan, outperforming a manually tuned reference curriculum from our internal development. These results show that a single parameter-efficient model can unify culturally and structurally distinct value-alignment benchmarks under a low-resource training budget."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="widanagamage-etal-2026-plurva">
<titleInfo>
<title>PlurVA-LLM-2026 Shared Task Track-1: Curriculum Fine-Tuning with Soft Labels and Positional Consistency for Cross-Cultural Value Alignment</title>
</titleInfo>
<name type="personal">
<namePart type="given">Amasha</namePart>
<namePart type="family">Widanagamage</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hansika</namePart>
<namePart type="family">Seneviratne</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Naween</namePart>
<namePart type="family">Fernando</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Dilushri</namePart>
<namePart type="family">Pavithra</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ruvan</namePart>
<namePart type="family">Weerasinghe</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Deshan</namePart>
<namePart type="family">Sumanathilaka</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>Proceedings of the First Workshop on Pluralistic Value Alignment of LLMs</title>
</titleInfo>
<name type="personal">
<namePart type="given">Deyi</namePart>
<namePart type="family">Xiong</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">António</namePart>
<namePart type="family">Branco</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hongming</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yue</namePart>
<namePart type="family">Dong</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Benyou</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Wenxuan</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Li</namePart>
<namePart type="family">Zhou</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jingting</namePart>
<namePart type="family">Zheng</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">979-8-89176-453-8</identifier>
</relatedItem>
<abstract>Pluralistic value alignment asks whether a language model can adapt its judgements to distinct, locally grounded value systems rather than reflecting a single aggregated notion of human preference. The PlurVA-LLM shared task evaluates this setting by unifying three independently constructed, culturally grounded multiple-choice value dilemmas across Chinese, Indonesian, and Sri Lankan datasets. We present a single-model submission based on Qwen3.5-4B with LoRA fine-tuning, trained only on the 1,359 labelled development instances provided by the task. Our method combines three ideas motivated directly by the task structure: soft-label supervision for Indonesian annotator-vote distributions, consistency regularisation across option permutations to reduce answer-position bias, and a staged multilingual curriculum with cross-validation-based iteration selection. On the hidden test set, our system achieves a macro-average accuracy of 0.8062, with 0.7788 on Chinese, 0.7139 on Indonesian, and 0.9260 on Sri Lankan, outperforming a manually tuned reference curriculum from our internal development. These results show that a single parameter-efficient model can unify culturally and structurally distinct value-alignment benchmarks under a low-resource training budget.</abstract>
<identifier type="citekey">widanagamage-etal-2026-plurva</identifier>
<location>
<url>https://aclanthology.org/2026.plurvallm-1.14/</url>
</location>
<part>
<date>2026-11</date>
<extent unit="page">
<start>187</start>
<end>196</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T PlurVA-LLM-2026 Shared Task Track-1: Curriculum Fine-Tuning with Soft Labels and Positional Consistency for Cross-Cultural Value Alignment
%A Widanagamage, Amasha
%A Seneviratne, Hansika
%A Fernando, Naween
%A Pavithra, Dilushri
%A Weerasinghe, Ruvan
%A Sumanathilaka, Deshan
%Y Xiong, Deyi
%Y Branco, António
%Y Zhang, Hongming
%Y Dong, Yue
%Y Wang, Benyou
%Y Zhang, Wenxuan
%Y Zhou, Li
%Y Zheng, Jingting
%S Proceedings of the First Workshop on Pluralistic Value Alignment of LLMs
%D 2026
%8 November
%I Association for Computational Linguistics
%C Hengqin, China
%@ 979-8-89176-453-8
%F widanagamage-etal-2026-plurva
%X Pluralistic value alignment asks whether a language model can adapt its judgements to distinct, locally grounded value systems rather than reflecting a single aggregated notion of human preference. The PlurVA-LLM shared task evaluates this setting by unifying three independently constructed, culturally grounded multiple-choice value dilemmas across Chinese, Indonesian, and Sri Lankan datasets. We present a single-model submission based on Qwen3.5-4B with LoRA fine-tuning, trained only on the 1,359 labelled development instances provided by the task. Our method combines three ideas motivated directly by the task structure: soft-label supervision for Indonesian annotator-vote distributions, consistency regularisation across option permutations to reduce answer-position bias, and a staged multilingual curriculum with cross-validation-based iteration selection. On the hidden test set, our system achieves a macro-average accuracy of 0.8062, with 0.7788 on Chinese, 0.7139 on Indonesian, and 0.9260 on Sri Lankan, outperforming a manually tuned reference curriculum from our internal development. These results show that a single parameter-efficient model can unify culturally and structurally distinct value-alignment benchmarks under a low-resource training budget.
%U https://aclanthology.org/2026.plurvallm-1.14/
%P 187-196
Markdown (Informal)
[PlurVA-LLM-2026 Shared Task Track-1: Curriculum Fine-Tuning with Soft Labels and Positional Consistency for Cross-Cultural Value Alignment](https://aclanthology.org/2026.plurvallm-1.14/) (Widanagamage et al., PlurVA-LLM 2026)
ACL