大语言模型中风险敏感决策的行为特征
Behavioural Signatures of Risk-Sensitive Decision-Making in Large Language Models
查看机构详情
- School of Computer Science, Wuhan University（武汉大学计算机科学学院）
- College of Computing and Data Science, Nanyang Technological University（南洋理工大学计算与数据科学学院）
机构由 AI 辅助整理，请以论文原文为准。
浏览论文内容
中文总结 AI 辅助
研究大语言模型在决策支持中风险敏感决策的行为特征，通过无限制德州扑克的多模型框架量化行为，发现前沿LLMs有稳定风险特征，在不同条件下表现各异，为审计交互式环境中风险敏感决策提供行为基础。
中文摘要 AI 辅助
随着大语言模型（LLMs）越来越多地用于决策支持，了解它们在不确定性下的选择是否表现出稳定且可解释的行为规律很重要。人类决策结合了相对持久的风险偏好和依赖上下文的调整，而基于LLM的决策系统中是否能观察到类似行为结构尚不清楚。本文使用基于无限制德州扑克的受控多模型框架研究此问题，行为通过参与度（衡量对不确定机会的自愿参与）和主动性（衡量翻牌前风险升级）来量化。在同质自博弈和异质混合模型交互中，前沿LLMs表现出稳定、特定模型的风险特征，形成从保守到激进的决策风格谱。这些特征在对手组成变化时基本保持稳健，最保守和最激进的模型在混合设置中差异更大。在全球风险压力和个人资源约束下，模型以结构化但异质的方式适应，从广泛的行为收缩到选择性降级和近乎不变的行为。这些发现表明LLMs不仅在基线风险倾向方面存在差异，在它们响应的风险信号和调整的灵活性方面也存在差异，为审计交互式环境中风险敏感决策提供了行为基础。
英文摘要
As large language models (LLMs) are increasingly used in decision support, it is important to understand whether their choices under uncertainty exhibit stable and interpretable behavioural regularities. Human decision-making combines relatively persistent risk preferences with context-dependent adjustment, yet it remains unclear whether analogous behavioural structure can be observed in LLM-based decision systems. Here we examine this question using a controlled multi-model framework based on no-limit Texas Hold'em, where behaviour is quantified by Participation, measuring voluntary engagement in uncertain opportunities, and Proactiveness, measuring pre-flop risk escalation. Across homogeneous self-play and heterogeneous mixed-model interactions, frontier LLMs exhibit stable, model-specific risk profiles, forming a spectrum from conservative to aggressive decision styles. These profiles remain largely robust under changing opponent composition, while the most conservative and most aggressive models diverge further in mixed settings. Under global risk pressure and personal resource constraint, models adapt in structured but heterogeneous ways, ranging from broad behavioural contraction to selective de-escalation and near-invariant behaviour. These findings suggest that LLMs differ not only in baseline risk disposition, but also in the risk signals they respond to and the flexibility with which they adjust, providing a behavioural basis for auditing risk-sensitive decision-making in interactive settings. Our code is publicly available at: https://github.com/XuankunRong/AgentTexasPoker.
