arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2503.06646econ.GNcs.CLq-fin.EC

评估和对齐LLMs中的人类经济风险偏好

Evaluating and Aligning Human Economic Risk Preferences in LLMs

The Hong Kong University of Science and Technology(香港科技大学)

Jiaxin Liu, Yixuan Tang, Yi Yang, Kar Yan Tam

AI总结:

本研究评估了LLMs在不同人物设定下的经济风险偏好,发现其在复杂经济决策任务中表现下降,并提出一种对齐方法以增强LLM对特定人物设定的风险偏好的遵循,从而提升其在风险相关应用中的经济理性。

摘要:

大型语言模型(LLMs)越来越多地应用于涉及风险评估的决策场景中,然而它们与人类经济理性的对齐程度仍不明确。在本研究中,我们调查了LLMs是否在不同人物设定下表现出与人类期望一致的风险偏好。具体而言,我们评估LLM生成的响应是否反映了基于个人设定的适当程度的风险厌恶或风险寻求行为。我们的结果表明,虽然LLMs在简化的、个性化的风险情境中做出合理的决策,但它们在更复杂的经济决策任务中表现下降。为了解决这个问题,我们提出了一种对齐方法,旨在增强LLM对特定人物设定的风险偏好的遵循。我们的方法提高了LLMs在风险相关应用中的经济理性,为更符合人类期望的AI决策迈出了一步。

英文摘要:

Large Language Models (LLMs) are increasingly used in decision-making scenarios that involve risk assessment, yet their alignment with human economic rationality remains unclear. In this study, we investigate whether LLMs exhibit risk preferences consistent with human expectations across different personas. Specifically, we assess whether LLM-generated responses reflect appropriate levels of risk aversion or risk-seeking behavior based on individual's persona. Our results reveal that while LLMs make reasonable decisions in simplified, personalized risk contexts, their performance declines in more complex economic decision-making tasks. To address this, we propose an alignment method designed to enhance LLM adherence to persona-specific risk preferences. Our approach improves the economic rationality of LLMs in risk-related applications, offering a step toward more human-aligned AI decision-making.

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。

↑