arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2502.11423cs.CL

探索个性化对话生成中的人物角色情感敏感性

Exploring Persona Sentiment Sensitivity in Personalized Dialogue Generation

Department of Artificial Intelligence, Chung-Ang University(人工智能系, Chung-Ang 大学)

Yonghyun Jun, Hwanhee Lee

AI总结:

本文通过大规模实验揭示用户人物角色情感极性对LLM个性化对话质量的影响,并提出结合轮次生成、画像排序和情感感知提示的对话生成方法,以提升个性化对话系统的稳健性与细腻度。

摘要:

随着将用户特定的人物角色(persona)集成到大语言模型(LLMs)中,个性化对话系统已取得显著进展。然而,尽管LLMs能够有效生成个性化回复,人物角色情感对对话质量的影响仍未得到充分探索。在本工作中,我们对使用一系列极化用户画像生成的对话进行了大规模分析。实验表明,涉及负向极化用户的对话往往会过度强调人物角色属性;相比之下,正向极化的画像生成的对话会选择性地融入人物角色信息,从而产生更流畅的交互。此外,我们发现情感较弱或中性的人物角色通常会产生质量较低的对话。基于这些发现,我们提出了一种对话生成方法,该方法通过将基于轮次的生成策略与画像排序机制以及情感感知提示相结合,显式地考虑人物角色极性。本研究为LLMs对人物角色情感的敏感性提供了新的见解,并为开发更稳健、更细腻的个性化对话系统提供了指导。

英文摘要:

Personalized dialogue systems have advanced considerably with the integration of user-specific personas into large language models (LLMs). However, while LLMs can effectively generate personalized responses, the influence of persona sentiment on dialogue quality remains underexplored. In this work, we conduct a large-scale analysis of dialogues generated using a range of polarized user profiles. Our experiments reveal that dialogues involving negatively polarized users tend to overemphasize persona attributes. In contrast, positively polarized profiles yield dialogues that selectively incorporate persona information, resulting in smoother interactions. Furthermore, we find that personas with weak or neutral sentiment generally produce lower-quality dialogues. Motivated by these findings, we propose a dialogue generation approach that explicitly accounts for persona polarity by combining a turn-based generation strategy with a profile ordering mechanism and sentiment-aware prompting. Our study provides new insights into the sensitivity of LLMs to persona sentiment and offers guidance for developing more robust and nuanced personalized dialogue systems.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑