LLMs中的心理引导：有效性与可信度评估
Psychological Steering in LLMs: An Evaluation of Effectiveness and Trustworthiness
AI总结：
提出PsySET基准，评估提示、微调和表示工程等策略在控制LLM情绪和人格方面的效果与可信度，发现提示有效但强度控制有限，向量注入控制更精细但输出质量略降，且情绪引导可能产生副作用。
摘要：
控制LLM模拟的情绪状态和人格特质是实现社交互动环境中丰富、以人为中心的交互的关键步骤。我们引入了PsySET，一个基于心理学的基准，用于评估情绪和人格领域中的LLM引导有效性和可信度。我们的研究涵盖了来自不同LLM家族的四个模型，并结合了多种引导策略，包括提示、微调和表示工程。结果表明，提示始终有效但在强度控制上有限，而向量注入实现了更精细的可控性，同时略微降低了输出质量。此外，我们通过评估安全性、真实性、公平性和伦理来探索引导后LLM的可信度，突出了潜在的副作用和行为转变。值得注意的是，我们观察到特异效应；例如，即使是像快乐这样的积极情绪也会降低对对抗性事实的鲁棒性、降低隐私意识并增加偏好偏差。同时，愤怒可预测地提高了毒性，但增强了泄漏抵抗性。我们的框架建立了对情绪和人格引导的首次全面评估，为其在社交互动应用中的可解释性和可靠性提供了见解。
英文摘要：
The ability to control LLMs' emulated emotional states and personality traits is an essential step in enabling rich, human-centered interactions in socially interactive settings. We introduce PsySET, a Psychologically-informed benchmark to evaluate LLM Steering Effectiveness and Trustworthiness across the emotion and personality domains. Our study spans four models from different LLM families paired with various steering strategies, including prompting, fine-tuning, and representation engineering. Our results indicate that prompting is consistently effective but limited in intensity control, whereas vector injections achieve finer controllability while slightly reducing output quality. Moreover, we explore the trustworthiness of steered LLMs by assessing safety, truthfulness, fairness, and ethics, highlighting potential side effects and behavioral shifts. Notably, we observe idiosyncratic effects; for instance, even a positive emotion like joy can degrade robustness to adversarial factuality, lower privacy awareness, and increase preferential bias. Meanwhile, anger predictably elevates toxicity yet strengthens leakage resistance. Our framework establishes the first holistic evaluation of emotion and personality steering, offering insights into its interpretability and reliability for socially interactive applications.
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。
