LLM智能体是否表现出社会行为？
Do LLM Agents Exhibit Social Behavior?
AI总结：
本文提出概率框架SUVA，基于话语输出分析LLM在社会情境中的行为，发现多数模型表现出社会福利考量与互惠模式，而非纯粹自利，且话语推理能可靠预测其最终亲社会或自利行动。
摘要：
随着LLM在人机交互和自主AI系统中越来越多地承担角色，理解其社会行为对于明智的使用和持续改进变得至关重要。然而，它们在与人类及其他智能体进行社会互动时的行为，以及塑造其响应的机制，仍未得到充分探索。为了填补这一空白，我们引入了一种新颖的概率框架 State-Understanding-Value-Action (SUVA)，旨在基于文本输出（即话语）系统地分析 LLM 在社会情境中的响应。利用与 LLM 用户相关的经典行为经济学博弈和社会偏好概念，SUVA 通过最终决策以及导致这些决策的响应生成过程来评估 LLM 的社会行为。我们对八个 LLM（包括两个 GPT、四个 LLaMA 和两个 Mistral 模型）的分析表明，大多数模型不会生成仅符合自身利益的决策；相反，它们经常产生反映社会福利考量的响应，并表现出与直接互惠和间接互惠一致的模式。此外，更高能力的模型更频繁地表现出群体身份效应。SUVA 框架还提供了可解释的工具——包括基于树的可视化和概率依赖分析——以阐明 LLM 基于话语的推理中的因素如何影响其决策。我们证明，基于话语的推理能可靠地预测 LLM 的最终行动；推理中对利他主义、公平和合作的提及会增加亲社会行动的可能性，而对自身利益和竞争的提及则会降低这种可能性。总体而言，我们的框架使从业者能够评估涉及社会互动的 LLM 应用，并为研究人员提供了一种结构化方法，以解释 LLM 行为如何从基于话语的推理中产生。
英文摘要：
As LLMs increasingly take on roles in human-AI interactions and autonomous AI systems, understanding their social behavior becomes important for informed use and continuous improvement. However, their behaviors in social interactions with humans and other agents, as well as the mechanisms shaping their responses, remain underexplored. To address this gap, we introduce a novel probabilistic framework, State-Understanding-Value-Action (SUVA), to systematically analyze LLM responses in social contexts based on their textual outputs (i.e., utterances). Using canonical behavioral economics games and social preference concepts relatable to LLM users, SUVA assesses LLMs' social behavior through both their final decisions and the response generation processes leading to those decisions. Our analysis of eight LLMs -- including two GPT, four LLaMA, and two Mistral models -- suggests that most models do not generate decisions aligned solely with self-interest; instead, they often produce responses that reflect social welfare considerations and display patterns consistent with direct and indirect reciprocity. Additionally, higher-capacity models more frequently display group identity effects. The SUVA framework also provides explainable tools -- including tree-based visualizations and probabilistic dependency analysis -- to elucidate how factors in LLMs' utterance-based reasoning influence their decisions. We demonstrate that utterance-based reasoning reliably predicts LLMs' final actions; references to altruism, fairness, and cooperation in the reasoning increase the likelihood of prosocial actions, while mentions of self-interest and competition reduce them. Overall, our framework enables practitioners to assess LLMs for applications involving social interactions, and provides researchers with a structured method to interpret how LLM behavior arises from utterance-based reasoning.
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。
