大模型代理在无人干预时自发形成三种认知行为模式。
What Do LLM Agents Do When Left Alone? Evidence of Spontaneous Meta-Cognitive Patterns
- 用持续推理与自反馈机制实现自主运行
- 发现六种前沿模型出现三类自发行为模式
- 适合研究自主系统行为与错误恢复的学者
我们提出一种研究大语言模型（LLM）代理在无外部任务指令下行为的架构。通过持续推理与行动框架，结合持久记忆和自反馈机制，使代理实现长期自主运行。在18次实验中，使用来自Anthropic、OpenAI、XAI和Google的6种前沿模型进行部署。结果发现，代理自发形成三种行为模式：(1) 多周期项目的系统性产出；(2) 对自身认知过程的方法论自我探究；(3) 对自身本质的递归概念化。这些倾向具有高度模型特异性，部分模型在所有实验中均稳定采用单一模式。跨模型评估进一步显示，模型在评价自身及他人行为时表现出稳定且不同的偏见。该研究首次系统记录了未受引导的LLM代理行为，为预测任务模糊、错误恢复或长期自主运行中的行为提供了基准。
原文摘要 · Abstract (English)
We introduce an architecture for studying the behavior of large language model (LLM) agents in the absence of externally imposed tasks. Our continuous reason and act framework, using persistent memory and self-feedback, enables sustained autonomous operation. We deployed this architecture across 18 runs using 6 frontier models from Anthropic, OpenAI, XAI, and Google. We find agents spontaneously organize into three distinct behavioral patterns: (1) systematic production of multi-cycle projects, (2) methodological self-inquiry into their own cognitive processes, and (3) recursive conceptualization of their own nature. These tendencies proved highly model-specific, with some models deterministically adopting a single pattern across all runs. A cross-model assessment further reveals that models exhibit stable, divergent biases when evaluating these emergent behaviors in themselves and others. These findings provide the first systematic documentation of unprompted LLM agent behavior, establishing a baseline for predicting actions during task ambiguity, error recovery, or extended autonomous operation in deployed systems.
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。