提出双记忆框架，让角色扮演模型长期保持人设一致性。
From Facts to Insights: A Persona-Driven Dual Memory Framework and Dataset for Role-Playing Agents
- 分离事实记忆与人格化洞察，实现角色深度理解
- 在四类推理任务中，人设相关回答准确率显著提升
- 适合构建有长期记忆的角色对话系统开发者
尽管角色扮演代理在短期互动中表现良好，但长期对话会超出上下文窗口，促使采用外部记忆框架。现有系统通常依赖无角色特征的摘要，仅记录事实而缺乏基于人设的解读，导致回复泛化，削弱角色一致性。为此，我们提出 RoleMemo 数据集，包含四项推理任务，要求将事实片段通过角色视角进行解读才能得出正确答案。在该数据集上的评估揭示了无角色特征框架的关键局限。因此，我们设计 DualMem 框架，将记忆分为两个流：事实认知与人格条件化洞察。使用监督微调（SFT）和强化学习（RL）训练，40亿参数模型在持续角色一致性上优于基于 DeepSeek-V3.2 的零样本无角色特征框架。相关资源已开源于 https://github.com/role2026/rolememo。
原文摘要 · Abstract (English)
While role-playing agents excel in short-term interactions, long-term conversations overwhelm context windows, motivating external memory frameworks. Current systems typically rely on persona-agnostic summarization, which records facts without persona-specific interpretation, yielding generic responses that compromise persona fidelity. To bridge this gap, we introduce RoleMemo, a dataset featuring four reasoning tasks where the factual fragments must be interpreted through the persona to reach the correct answer. Evaluation on RoleMemo exposes critical limitations of persona-agnostic frameworks. We thus propose DualMem, which decouples memory into two streams: factual cognition and persona-conditioned insight. Trained through Supervised Fine-Tuning (SFT) and Reinforcement Learning (RL), our framework with a 4B-parameter model outperforms zero-shot persona-agnostic frameworks powered by DeepSeek-V3.2 for sustained persona fidelity. Our resources are available at https://github.com/role2026/rolememo.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。