arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2610.18541cs.CL
收藏

共情并非事实性与可靠性下降的替罪羊

Empathy Is Not the Scapegoat for Declines in Factuality and Reliability

INSAIT, Sofia University “St. Kliment Ohridski”(索菲亚大学“圣克利门特·奥赫里德斯基” INSAIT);University of British Columbia(不列颠哥伦比亚大学)

Letian Shi, Kanwal Mehreen, Deheng Zhang, Yuxia Wang

AI总结:

该研究针对语言模型温暖性与可靠性的关联问题,通过多层面分析发现任务对齐的自适应语气可兼顾共情性与可靠性,还引入了人工验证的共情评估器,证明模型可兼具社交支持性与可靠性。

摘要:

语言模型越来越被期望在用户处于压力或不确定情境时提供支持,此时温暖与共情是理想特质。尽管近期研究将更具温暖性的模型与可靠性降低关联起来,但我们表明该结论并不完整。过往研究通常主要基于共情对话进行训练,这使得难以将语气效应与SFT(监督微调)引发的分布偏移分离开来。我们从模型和评估两个层面重新审视该问题,研究提示、SFT和强化学习(RL)在诱导温暖行为中的作用,并使用包含11个数据集的基准套件以及EMOFACT在情感语境下全面评估能力保留情况。在多个模型家族中,我们发现温暖效应具有任务和模型依赖性,而与任务对齐的自适应语气可提升共情性与针对性可靠性,尽管仍存在一些能力权衡。我们还引入了经人工验证的共情评估器,专门用于评估模型的共情响应。我们的结果表明,模型可在社交互动中更具支持性,同时保持可靠性。

英文摘要:

Language models are increasingly expected to support users in stressful or uncertain situations, where warmth and empathy are desirable. While recent work associates warmer models with reduced reliability, we show that this conclusion is incomplete. Prior studies often train primarily on empathetic dialogue, making it difficult to separate tone effects from SFT-induced distribution shift. We revisit this issue at both the model and evaluation levels. We study prompting, SFT, and RL for inducing warm behavior, and comprehensively evaluate capability preservation using an 11-dataset benchmark suite and EMOFACT under emotional contexts. Across multiple model families, we find that warmth effects are task- and model-dependent, while task-aligned adaptive tone can improve empathy and targeted reliability, although some capability trade-offs remain. We further introduce human-validated empathy evaluators specifically developed to assess empathetic model responses. Our results suggest that models can be more supportive in social interactions while remaining reliable.

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。

↑