@inproceedings{lai-etal-2026-rolecde,
title = "{R}ole{CDE}: Benchmarking and Mitigating Role{--}Alignment Trade-offs in Role-Playing Agents",
author = "Lai, Huayi and
Song, Shichao and
Niu, Simin and
Wang, Hanyu and
Yang, Jiawei and
Wang, Zhouxing and
Yin, Zhiqiang and
Liang, Xun",
editor = "Liakata, Maria and
Moreira, Viviane P. and
Zhang, Jiajun and
Jurgens, David",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {ACL} 2026",
month = jul,
year = "2026",
address = "San Diego, California, United States",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-acl.106/",
pages = "2226--2248",
ISBN = "979-8-89176-395-1",
abstract = "Role-playing agents(RPAs) are widely used to steer large language models(LLMs) toward role-consistent behavior, yet existing benchmarks mainly evaluate surface-level fidelity and offer limited insight into decision making under role{--}alignment value conflicts. To address this gap, we introduce RoleCDE, the first benchmark designed to evaluate RPAs under structured conflicts between role-specific values and alignment-oriented constraints. RoleCDE formulates role-aware decision making as cognitive dilemma scenarios, jointly evaluating role{--}scenario grounding, value conflict resolution, and decision tendencies. The benchmark is constructed at scale, covering approximately 8k diverse role profiles and scenarios and nearly 240k dilemma instances across three difficulty levels and eight role categories. Evaluation of several mainstream LLMs reveals a ``Role Value Decoupling'' phenomenon, where agents systematically default to alignment- and morality-consistent decisions rather than role-specific values when the two conflict, even under explicit role conditioning. This behavior is largely invariant to dilemma difficulty but varies substantially across role categories. We further show that RoleCDE-based fine-tuning effectively mitigates this decoupling by improving value trade-off reasoning, while preserving general role-playing fidelity and general reasoning performance. Code is available at: \url{https://github.com/rabbitrose/RoleCDE}."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="lai-etal-2026-rolecde">
<titleInfo>
<title>RoleCDE: Benchmarking and Mitigating Role–Alignment Trade-offs in Role-Playing Agents</title>
</titleInfo>
<name type="personal">
<namePart type="given">Huayi</namePart>
<namePart type="family">Lai</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shichao</namePart>
<namePart type="family">Song</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Simin</namePart>
<namePart type="family">Niu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hanyu</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jiawei</namePart>
<namePart type="family">Yang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhouxing</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhiqiang</namePart>
<namePart type="family">Yin</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xun</namePart>
<namePart type="family">Liang</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>Findings of the Association for Computational Linguistics: ACL 2026</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">979-8-89176-395-1</identifier>
</relatedItem>
<abstract>Role-playing agents(RPAs) are widely used to steer large language models(LLMs) toward role-consistent behavior, yet existing benchmarks mainly evaluate surface-level fidelity and offer limited insight into decision making under role–alignment value conflicts. To address this gap, we introduce RoleCDE, the first benchmark designed to evaluate RPAs under structured conflicts between role-specific values and alignment-oriented constraints. RoleCDE formulates role-aware decision making as cognitive dilemma scenarios, jointly evaluating role–scenario grounding, value conflict resolution, and decision tendencies. The benchmark is constructed at scale, covering approximately 8k diverse role profiles and scenarios and nearly 240k dilemma instances across three difficulty levels and eight role categories. Evaluation of several mainstream LLMs reveals a “Role Value Decoupling” phenomenon, where agents systematically default to alignment- and morality-consistent decisions rather than role-specific values when the two conflict, even under explicit role conditioning. This behavior is largely invariant to dilemma difficulty but varies substantially across role categories. We further show that RoleCDE-based fine-tuning effectively mitigates this decoupling by improving value trade-off reasoning, while preserving general role-playing fidelity and general reasoning performance. Code is available at: https://github.com/rabbitrose/RoleCDE.</abstract>
<identifier type="citekey">lai-etal-2026-rolecde</identifier>
<location>
<url>https://aclanthology.org/2026.findings-acl.106/</url>
</location>
<part>
<date>2026-07</date>
<extent unit="page">
<start>2226</start>
<end>2248</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T RoleCDE: Benchmarking and Mitigating Role–Alignment Trade-offs in Role-Playing Agents
%A Lai, Huayi
%A Song, Shichao
%A Niu, Simin
%A Wang, Hanyu
%A Yang, Jiawei
%A Wang, Zhouxing
%A Yin, Zhiqiang
%A Liang, Xun
%Y Liakata, Maria
%Y Moreira, Viviane P.
%Y Zhang, Jiajun
%Y Jurgens, David
%S Findings of the Association for Computational Linguistics: ACL 2026
%D 2026
%8 July
%I Association for Computational Linguistics
%C San Diego, California, United States
%@ 979-8-89176-395-1
%F lai-etal-2026-rolecde
%X Role-playing agents(RPAs) are widely used to steer large language models(LLMs) toward role-consistent behavior, yet existing benchmarks mainly evaluate surface-level fidelity and offer limited insight into decision making under role–alignment value conflicts. To address this gap, we introduce RoleCDE, the first benchmark designed to evaluate RPAs under structured conflicts between role-specific values and alignment-oriented constraints. RoleCDE formulates role-aware decision making as cognitive dilemma scenarios, jointly evaluating role–scenario grounding, value conflict resolution, and decision tendencies. The benchmark is constructed at scale, covering approximately 8k diverse role profiles and scenarios and nearly 240k dilemma instances across three difficulty levels and eight role categories. Evaluation of several mainstream LLMs reveals a “Role Value Decoupling” phenomenon, where agents systematically default to alignment- and morality-consistent decisions rather than role-specific values when the two conflict, even under explicit role conditioning. This behavior is largely invariant to dilemma difficulty but varies substantially across role categories. We further show that RoleCDE-based fine-tuning effectively mitigates this decoupling by improving value trade-off reasoning, while preserving general role-playing fidelity and general reasoning performance. Code is available at: https://github.com/rabbitrose/RoleCDE.
%U https://aclanthology.org/2026.findings-acl.106/
%P 2226-2248
Markdown (Informal)
[RoleCDE: Benchmarking and Mitigating Role–Alignment Trade-offs in Role-Playing Agents](https://aclanthology.org/2026.findings-acl.106/) (Lai et al., Findings 2026)
ACL
- Huayi Lai, Shichao Song, Simin Niu, Hanyu Wang, Jiawei Yang, Zhouxing Wang, Zhiqiang Yin, and Xun Liang. 2026. RoleCDE: Benchmarking and Mitigating Role–Alignment Trade-offs in Role-Playing Agents. In Findings of the Association for Computational Linguistics: ACL 2026, pages 2226–2248, San Diego, California, United States. Association for Computational Linguistics.