paper-with-me

홈 › Papers

RoleCDE:Benchmarking and Mitigating Role-Alignment Trade-offs in Role-Playing Agents

2026-06-01 · Huayi Lai, Shichao Song, Simin Niu, Hanyu Wang, Jiawei Yang, Zhouxing Wang, Zhiqiang Yin, Xun Liang arxiv

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 24k 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.

📄 PDF Abstract BibTeX arXiv:2606.01552

Code (0)

등록된 구현이 없습니다.

Tasks

Decision Making

Similar Papers 제목 키워드 기반

Debiasing Reward Models via Causally Motivated Inference-Time Intervention

2026-04-30 · Kazutoshi Shinoda, Kosuke Nishida, Kyosuke Nishida arxiv

Reward models (RMs) play a central role in aligning large language models (LLMs) with human preferences. However, RMs are often sensitive to spurious features such as response length. Existing inference-time approaches f…

CTIGuardian: A Few-Shot Framework for Mitigating Privacy Leakage in Fine-Tuned LLMs

2025-12-15 · Shashie Dilhara Batan Arachchige, Benjamin Zi Hao Zhao, Hassan Jameel Asghar, Dinusha Vatsalan 외 arxiv

Large Language Models (LLMs) are often fine-tuned to adapt their general-purpose knowledge to specific tasks and domains such as cyber threat intelligence (CTI). Fine-tuning is mostly done through proprietary datasets th…

Navigating the Alignment-Calibration Trade-off: A Pareto-Superior Frontier via Model Merging

2025-10-20 · Tiancheng Hu, Benjamin Minixhofer, Nigel Collier arxiv

The "alignment tax" of post-training is typically framed as a drop in task accuracy. We show it also involves a severe loss of calibration, making models overconfident, less reliable, and model outputs less diverse. We s…

Mitigating the Alignment Tax of RLHF

2023-09-12 · Yong Lin, Hangyu Lin, Wei Xiong, Shizhe Diao 외

LLMs acquire a wide range of abilities during pre-training, but aligning LLMs under Reinforcement Learning with Human Feedback (RLHF) can lead to forgetting pretrained abilities, which is also known as the alignment tax.…

Common Sense ReasoningContinual Learning

DEBATE: A Large-Scale Benchmark for Evaluating Opinion Dynamics in Role-Playing LLM Agents

2025-10-29 · Yun-Shiuan Chuang, Ruixuan Tu, Chengtao Dai, You Li 외 arxiv

Accurately modeling opinion change through social interactions is crucial for understanding and mitigating polarization, misinformation, and societal conflict. Recent work simulates opinion dynamics with role-playing LLM…