研究人员开发了“Agent Capsules”，一个旨在优化多智能体大型语言模型（LLM）管道的自适应运行时系统。该系统解决了合并智能体调用以节省 token 与潜在质量下降之间的权衡问题。Agent Capsules 根据经验质量约束动态选择复合执行策略，确保性能与 LangGraph 和 DSPy 等现有方法相当或更优，同时显著减少 token 使用量。 AI
影响 引入了一种新颖的运行时系统，用于优化 LLM 智能体管道，可能降低运营成本并提高效率。
排序理由 学术论文，详细介绍了用于优化多智能体 LLM 管道的新框架。
AI 生成摘要 · Google Gemini · 来自 2 个来源。 我们如何撰写摘要 →
arXiv:2605.00410v1 Announce Type: new Abstract: A multi-agent pipeline with N agents typically issues N LLM calls per run. Merging agents into fewer calls (compound execution) promises token savings, but naively merged calls silently degrade quality through tool loss and prompt c…
A multi-agent pipeline with N agents typically issues N LLM calls per run. Merging agents into fewer calls (compound execution) promises token savings, but naively merged calls silently degrade quality through tool loss and prompt compression. We present Agent Capsules, an adapti…