研究大模型协作时如何自发形成一致意见，揭示无角色提示下的群体智能。
Emergent Convergence in Multi-Agent LLM Annotation
- 通过7500次多轮对话模拟，追踪模型间互动与最终标注演化
- 模型在词汇与语义层面实现收敛，输出嵌入维度下降体现语义压缩
- 适用于研究大模型协作机制或评估群体智能的科研人员
大型语言模型（LLMs）越来越多地应用于协作场景，但对其作为黑箱代理时的协调机制知之甚少。我们模拟了7500次多代理、多轮次的归纳编码任务，生成超过12.5万条话语，记录最终标注及其交互历史。引入过程级指标：代码稳定性、语义自一致性、词汇置信度、情感与收敛度量，以追踪协调动态。为深入探查对齐信号，分析输出嵌入的演化几何结构，发现内在维度随轮次递减，表明语义压缩。结果表明，尽管无显式角色提示，模型群体仍实现词汇与语义收敛，表现出非对称影响力模式和类谈判行为。本工作展示了黑箱交互分析如何揭示涌现的协调策略，为内部探针式可解释性方法提供了可扩展的补充。
原文摘要 · Abstract (English)
Large language models (LLMs) are increasingly deployed in collaborative settings, yet little is known about how they coordinate when treated as black-box agents. We simulate 7500 multi-agent, multi-round discussions in an inductive coding task, generating over 125000 utterances that capture both final annotations and their interactional histories. We introduce process-level metrics: code stability, semantic self-consistency, and lexical confidence alongside sentiment and convergence measures, to track coordination dynamics. To probe deeper alignment signals, we analyze the evolving geometry of output embeddings, showing that intrinsic dimensionality declines over rounds, suggesting semantic compression. The results reveal that LLM groups converge lexically and semantically, develop asymmetric influence patterns, and exhibit negotiation-like behaviors despite the absence of explicit role prompting. This work demonstrates how black-box interaction analysis can surface emergent coordination strategies, offering a scalable complement to internal probe-based interpretability methods.
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。