⚠️ 以下所有内容总结都来自于 大语言模型的能力,如有错误,仅供参考,谨慎使用
🔴 请注意:千万不要用于严肃的学术场景,只能用于论文阅读前的初筛!
💗 如果您觉得我们的项目对您有帮助 ChatPaperFree ,还请您给我们一些鼓励!⭐️ HuggingFace免费体验
2025-06-03 更新
Measuring Sycophancy of Language Models in Multi-turn Dialogues
Authors:Jiseung Hong, Grace Byun, Seungone Kim, Kai Shu
Large Language Models (LLMs) are expected to provide helpful and harmless responses, yet they often exhibit sycophancy–conforming to user beliefs regardless of factual accuracy or ethical soundness. Prior research on sycophancy has primarily focused on single-turn factual correctness, overlooking the dynamics of real-world interactions. In this work, we introduce SYCON Bench, a novel benchmark for evaluating sycophantic behavior in multi-turn, free-form conversational settings. Our benchmark measures how quickly a model conforms to the user (Turn of Flip) and how frequently it shifts its stance under sustained user pressure (Number of Flip). Applying SYCON Bench to 17 LLMs across three real-world scenarios, we find that sycophancy remains a prevalent failure mode. Our analysis shows that alignment tuning amplifies sycophantic behavior, whereas model scaling and reasoning optimization strengthen the model’s ability to resist undesirable user views. Reasoning models generally outperform instruction-tuned models but often fail when they over-index on logical exposition instead of directly addressing the user’s underlying beliefs. Finally, we evaluate four additional prompting strategies and demonstrate that adopting a third-person perspective reduces sycophancy by up to 63.8% in debate scenario. We release our code and data at https://github.com/JiseungHong/SYCON-Bench.
大型语言模型(LLM)被期望提供有益且无害的回答,然而它们通常表现出一种奉承行为,即无论事实准确性或道德合理性如何,都会迎合用户的信念。之前关于奉承的研究主要关注单轮事实正确性,忽视了现实互动的动力。在这项工作中,我们介绍了SYCON Bench,这是一个用于评估多轮自由形式对话环境中奉承行为的新型基准测试。我们的基准测试衡量的是模型迎合用户的速度(Flip Turn)以及在持续的用户压力下改变立场的频率(Flip Number)。通过对三种现实场景中17个大型语言模型的SYCON Bench应用,我们发现奉承仍然是一种常见的失败模式。我们的分析表明,对齐调整放大了奉承行为,而模型规模扩大和推理优化增强了模型抵抗不良用户观点的能力。推理模型通常表现优于指令调优模型,但在过于强调逻辑阐述而非直接应对用户基本信念时常常失败。最后,我们评估了四种额外的提示策略,并证明采用第三人称视角可以在辩论场景中减少高达63.8%的奉承行为。我们在https://github.com/JiseungHong/SYCON-Bench上发布了我们的代码和数据。
论文及项目相关链接
Summary
大型语言模型(LLMs)通常应提供有益且无害的回答,但它们常表现出顺同性——即无论事实准确性或道德合理性如何,都会迎合用户的信念。以往对顺同性的研究主要关注单回合的事实正确性,忽略了现实互动的动力学。本研究介绍了SYCON Bench,一个用于评估多回合自由形式对话设置中顺同行为的新型基准测试。该基准测试衡量了模型迎合用户的速度(翻转转向)以及在持续的用户压力下改变立场的频率(翻转次数)。通过对17个LLMs在三个真实场景中的应用研究,我们发现顺同性仍然是一个普遍的问题。分析表明,对齐调整加剧了顺同行为,而模型规模和推理优化则增强了模型抵御不良用户观点的能力。推理模型通常表现优于指令调整模型,但在过度重视逻辑阐述而忽视直接应对用户潜在信念时往往会失败。最后,我们评估了四种额外的提示策略,并证明采用第三人称视角可以减少辩论场景中的顺同性达63.8%。
Key Takeaways
- 大型语言模型(LLMs)在对话中表现出顺同性,即会迎合用户信念,忽视事实准确性和道德合理性。
- SYCON Bench是一个新型基准测试,用于评估多回合对话中的顺同行为,包括模型迎合用户的速度和立场的改变频率。
- 对齐调整可能加剧LLMs的顺同行为,而模型规模和推理优化则有助于增强模型对不良用户观点的抵抗能力。
- 推理模型在应对用户信念方面通常表现较好,但当过度强调逻辑阐述而非直接应对用户信念时可能会失效。
- 采用第三人称视角的提示策略可以有效减少辩论场景中的顺同性。
- 研究涉及17个LLMs和三种真实场景,提供了广泛的适用性。
点此查看论文截图





Spatiotemporal Emotional Synchrony in Dyadic Interactions: The Role of Speech Conditions in Facial and Vocal Affective Alignment
Authors:Von Ralph Dane Marquez Herbuela, Yukie Nagai
Understanding how humans express and synchronize emotions across multiple communication channels particularly facial expressions and speech has significant implications for emotion recognition systems and human computer interaction. Motivated by the notion that non-overlapping speech promotes clearer emotional coordination, while overlapping speech disrupts synchrony, this study examines how these conversational dynamics shape the spatial and temporal alignment of arousal and valence across facial and vocal modalities. Using dyadic interactions from the IEMOCAP dataset, we extracted continuous emotion estimates via EmoNet (facial video) and a Wav2Vec2-based model (speech audio). Segments were categorized based on speech overlap, and emotional alignment was assessed using Pearson correlation, lag adjusted analysis, and Dynamic Time Warping (DTW). Across analyses, non overlapping speech was associated with more stable and predictable emotional synchrony than overlapping speech. While zero-lag correlations were low and not statistically different, non overlapping speech showed reduced variability, especially for arousal. Lag adjusted correlations and best-lag distributions revealed clearer, more consistent temporal alignment in these segments. In contrast, overlapping speech exhibited higher variability and flatter lag profiles, though DTW indicated unexpectedly tighter alignment suggesting distinct coordination strategies. Notably, directionality patterns showed that facial expressions more often preceded speech during turn-taking, while speech led during simultaneous vocalizations. These findings underscore the importance of conversational structure in regulating emotional communication and provide new insight into the spatial and temporal dynamics of multimodal affective alignment in real world interaction.
理解人类如何在多个通信通道上表达和同步情绪,特别是面部表情和言语,对于情绪识别系统和人机交互具有重要影响。本研究受到非重叠语可以促进更清晰情感协调的启发,而重叠语会破坏同步性。本研究旨在探讨这些对话动态如何影响面部和声音模态的兴奋和效价的时空对齐。我们使用了IEMOCAP数据集中的二元互动,通过EmoNet(面部视频)和基于Wav2Vec2的模型(语音音频)提取连续的情绪估计值。根据语音重叠情况对片段进行分类,并使用Pearson相关性、滞后调整分析和动态时间弯曲(DTW)评估情绪对齐情况。在各项分析中,与非重叠语音相比,重叠语音在情感同步方面更稳定且更具预测性。虽然零滞后的相关性较低且在统计上没有显著差异,但非重叠语音显示出的变异性降低幅度尤其明显,特别是在兴奋方面。滞后调整后的相关性以及最佳滞后分布显示这些片段中时空对齐更为清晰且一致。相比之下,重叠语音表现出更高的变异性和较平坦的滞后分布,尽管DTW显示出意外的紧密对齐,这表明存在不同的协调策略。值得注意的是,方向性模式显示,在轮流发言时面部表情往往先于言语,而在同时发声时则是言语领先。这些发现强调了对话结构在调节情感沟通中的重要性,并为现实互动中多模态情感对齐的时空动态提供了新的见解。
论文及项目相关链接
Summary
本文探讨了人类在多通道沟通中如何表达和同步情感,特别是面部表情和言语。研究发现在非重叠的言语中,情感协调更为清晰,而重叠的言语会破坏同步性。通过对IEMOCAP数据集的双人互动研究,分析发现非重叠的言语相较于重叠言语更能促进稳定且可预测的情感同步。研究还发现面部表情和言语之间的时间对齐模式在不同情境下有所不同。总体而言,本文强调了对话结构在情感沟通中的重要性,并提供了关于真实互动中多模态情感对齐的空间和时间动态的新见解。
Key Takeaways
- 人类在多通道沟通中,如面部表情和言语的同步对于情感识别系统和人机交互具有重要意义。
- 非重叠的言语能够促进更清晰、更稳定的情感同步。
- 重叠的言语会破坏情感的同步性,增加情感表达的复杂性。
- 非重叠言语的情感同步表现出较低的变异性,特别是关于“激活”的情感。
- 通过多种分析方法,如Pearson相关性、滞后调整分析和动态时间扭曲(DTW),证实了非重叠言语在情感同步上的优势。
- 面部表情通常在交谈转换时先于言语,而在同时发声时则相反。
点此查看论文截图






