arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2603.19282cs.CLcs.AI
收藏

独立代理大语言模型中的框架效应:跨家族行为分析

Framing Effects in Independent-Agent Large Language Models: A Cross-Family Behavioral Analysis

Northeastern University(东北大学);Beijing Institute of Technology(北京理工大学)

Zice Wang, Zhenyu Zhang

AI总结:

研究探讨了在非交互多代理部署中,提示框架如何影响独立代理在阈值投票任务中的决策,发现框架效应显著影响选择分布,倾向于风险规避选项,揭示了非交互多代理部署中的重要偏差源。

摘要:

在许多现实应用中,大语言模型(LLMs)作为独立代理运作,缺乏交互,从而限制了协调。在此设置中,我们研究了提示框架如何影响阈值投票任务中的决策,涉及个体-群体利益冲突。两个逻辑等价但框架不同的提示在不同LLM家族中进行了隔离测试。结果表明,提示框架显著影响选择分布,通常将偏好转向风险规避选项。表面语言线索甚至可以超越逻辑等价的表达。这表明观察到的行为反映了一种倾向于工具理性而非合作理性倾向,当成功需要承担风险时。研究结果突显了框架效应作为非交互多代理LLM部署中的重要偏差源,为对齐和提示设计提供了启示。

英文摘要:

In many real-world applications, large language models (LLMs) operate as independent agents without interaction, thereby limiting coordination. In this setting, we examine how prompt framing influences decisions in a threshold voting task involving individual-group interest conflict. Two logically equivalent prompts with different framings were tested across diverse LLM families under isolated trials. Results show that prompt framing significantly influences choice distributions, often shifting preferences toward risk-averse options. Surface linguistic cues can even override logically equivalent formulations. This suggests that observed behavior reflects a tendency consistent with a preference for instrumental rather than cooperative rationality when success requires risk-bearing. The findings highlight framing effects as a significant bias source in non-interacting multi-agent LLM deployments, informing alignment and prompt design.

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。

↑