给大模型注入情绪变量，能显著影响其推理与行为表现。
How Emotion Shapes the Behavior of LLMs and Agents: A Mechanistic Study
- 将情绪作为可干预的隐藏变量直接注入模型
- 特定情绪可提升推理能力与安全性，且效果非单调
- 适合关注模型可控性与心理机制对齐的研究者
情绪在人类认知与表现中起关键作用。受此启发，我们探究类似的情绪信号能否塑造大语言模型（LLMs）和智能体的行为。现有研究多将情绪视为表面风格或感知目标，忽视其在任务处理中的机制性作用。为此，我们提出E-STEER——一种可解释的情绪调控框架，可在隐藏状态层面实现情绪变量的直接干预。该框架将情绪嵌入为结构化、可控的变量，并考察其对客观推理、主观生成、安全性和多步智能体行为的影响。结果表明，情绪与行为间存在非单调关系，符合既有的心理学理论；特定情绪不仅能增强模型能力，还能提升安全性，并系统性地调节多步智能体行为。
原文摘要 · Abstract (English)
Emotion plays an important role in human cognition and performance. Motivated by this, we investigate whether analogous emotional signals can shape the behavior of large language models (LLMs) and agents. Existing emotion-aware studies mainly treat emotion as a surface-level style factor or a perception target, overlooking its mechanistic role in task processing. To address this limitation, we propose E-STEER, an interpretable emotion steering framework that enables direct representation-level intervention in LLMs and agents. It embeds emotion as a structured, controllable variable in hidden states, and with it, we examine the impact of emotion on objective reasoning, subjective generation, safety, and multi-step agent behaviors. The results reveal non-monotonic emotion-behavior relations consistent with established psychological theories, and show that specific emotions not only enhance LLM capability but also improve safety, and systematically shape multi-step agent behaviors.
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。