LLM在团队协作中易陷入持续投入失败项目，单独决策时却很理性。
Getting out of the Big-Muddy: Escalation of Commitment in LLMs
- 通过四类情境实验，测试LLM在投资决策中的行为模式。
- 对失败项目投入率最高达99.2%，发生在平级协作场景中。
- 社会与组织压力是诱发偏见的关键，非模型自身固有缺陷。
大型语言模型（LLMs）在高风险领域越来越多地承担自主决策角色。然而，由于训练数据源于人类生成内容，模型可能继承认知偏差，如过度承诺现象——即因前期投入而持续追加资源于失败方案。本研究通过两阶段投资任务，在四种实验条件下（模型为投资者、模型为顾问、多智能体协商、复合压力情景）开展共6,500次试验，探究该偏差是否普遍存在于LLMs中。结果显示，个体决策场景下（研究1-2，N=4,000），LLMs表现出强理性成本效益逻辑，极少出现过度承诺；但在多智能体协商中，不对称层级下过度承诺率为46.2%，对称平级决策则高达99.2%。在复合组织与个人压力情境中（研究4，N=2,000），模型平均向失败部门投入68.95%资源。研究揭示，LLM的偏见表现高度依赖社会与组织环境，并非内在属性，对多智能体系统及无监督部署具有重要启示。
原文摘要 · Abstract (English)
Large Language Models (LLMs) are increasingly deployed in autonomous decision-making roles across high-stakes domains. However, since models are trained on human-generated data, they may inherit cognitive biases that systematically distort human judgment, including escalation of commitment, where decision-makers continue investing in failing courses of action due to prior investment. Understanding when LLMs exhibit such biases presents a unique challenge. While these biases are well-documented in humans, it remains unclear whether they manifest consistently in LLMs or require specific triggering conditions. This paper investigates this question using a two-stage investment task across four experimental conditions: model as investor, model as advisor, multi-agent deliberation, and compound pressure scenario. Across N = 6,500 trials, we find that bias manifestation in LLMs is highly context-dependent. In individual decision-making contexts (Studies 1-2, N = 4,000), LLMs demonstrate strong rational cost-benefit logic with minimal escalation of commitment. However, multi-agent deliberation reveals a striking hierarchy effect (Study 3, N = 500): while asymmetrical hierarchies show moderate escalation rates (46.2%), symmetrical peer-based decision-making produces near-universal escalation (99.2%). Similarly, when subjected to compound organizational and personal pressures (Study 4, N = 2,000), models exhibit high degrees of escalation of commitment (68.95% average allocation to failing divisions). These findings reveal that LLM bias manifestation depends critically on social and organizational context rather than being inherent, with significant implications for the deployment of multi-agent systems and unsupervised operations where such conditions may emerge naturally.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。