诱导情绪会影响大语言模型在序列决策中的行为吗？

中文摘要 AI 辅助

随着大语言模型（LLMs）越来越多地在高风险领域中作为自主智能体部署，了解可能调节其决策的上下文因素变得至关重要。虽然LLMs经过训练以感知并与用户情绪产生共鸣，但诱导情绪是否会影响其序列决策仍不清楚。我们使用爱荷华赌博任务（IGT）（一种研究不确定性下决策的经典心理学范式）并结合基于想象的情绪诱导程序来研究这个问题。首先通过确认LLMs能从上下文中感知强烈且可区分的情绪，以及LLM智能体能以类似人类的速度从序列交互中学习，验证了该范式的可行性。在经过验证的设置下，我们发现，与人类不同，平均而言诱导情绪不会显著影响LLM智能体的决策动态。然而，愤怒的影响是有条件的：诱导愤怒会使LLM智能体对错误决策的惩罚不太敏感，并且在游戏早期，愤怒会降低探索，使决策过早锁定在少数选择上。这些发现揭示了诱导情绪对LLM决策与人类行为相比的细微但不同的影响，并为未来关于LLM智能体情感调节的研究提供了一种工具。

英文摘要

As Large Language Models (LLMs) are increasingly deployed as autonomous agents in high-stakes domains, understanding contextual factors that may modulate their decision-making becomes critical. While LLMs are trained to perceive and resonate with users' emotions, it remains unclear whether induced emotion can influence their sequential decision-making. We investigate this question using the Iowa Gambling Task (IGT), a classic psychological paradigm for studying decision-making under uncertainty, combined with an imagination-based emotion induction procedure. We first validate the feasibility of this paradigm by confirming that LLMs can sense strong, distinguishable emotions from context and that LLM agents can learn from sequential interactions in a human-like pace. With the validated setup, we find that, different from humans, induced emotion does not significantly bias the decision dynamics of LLM agents on average. However, the effects of anger are conditioned: inducing anger makes LLM agents less sensitive to penalties for bad decisions, and in early stages of the game, anger can lower exploration, locking decisions into a few choices early. These findings reveal the subtle yet distinct effects of induced emotion on LLM decision-making compared to human behavior, and provide a tool for future research on affective modulation of LLM agents.

发表机构

• MBZUAI（穆罕默德·本·扎耶德人工智能大学）
• Texas A&M University（德州农工大学）
• National University of Singapore（新加坡国立大学）
• UCLA（加州大学洛杉矶分校）
• University of Texas at Austin（德克萨斯大学奥斯汀分校）
• Mass General Hospital（麻省总医院）
• Harvard Medical School（哈佛医学院）

机构由 AI 辅助整理，请以论文原文为准。