知识卡片

结构图卡 · 8697.b · 生成模型:GPT-6

智能体系统的执行边界

理解等级:中级

结构图:

classDiagram
 class Orchestrator {
 state
 execute()
 record()
 }
 class Model {
 proposeAction()
 }
 class Tool {
 validateArguments()
 run()
 }
 class DataStore {
 retrieve()
 }
 class Verifier {
 check()
 }
 class Guardrail {
 allow()
 }
 Orchestrator --> Model : 提供状态
 Orchestrator --> Tool : 受控执行
 Orchestrator --> DataStore : 检索
 Orchestrator --> Verifier : 任务核查
 Orchestrator --> Guardrail : 边界检查

模型提出调用,运行时承担执行与状态责任;失败反馈进入下一轮,不能由模型自称执行成功。

这张卡片看不懂?

已有反馈 · 0 条

    还没有人提交反馈。

    在整书笔记中的位置

    原书出处 · 1 处

    • · 查看原书依据

      Figure 10-5 shows a canonical agentic system and how its components interact. Figure 10-5. A production-grade agentic system Let’s explore each of these types. Models Language models are the backbone of agentic systems, responsible for their autonomous nature and problem-solving capabilities. A single agentic system could be composed of multiple language models, with each model playing a distinct role. For example, you can build an

    参考来源

    - 位置:《Designing Large Language Model Applications》第10章《外部工具与智能体》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Designing Large Language Model Applications/_epub-src/OEBPS/ch10.html) - 结论依据:摘录中关于智能体系统的执行边界的定义、比较或条件支撑卡片;原书表述不严谨的范围已在覆盖记录逐项限定,不把案例结果外推为普遍结论。 - 原始内容:Figure 10-5 shows a canonical agentic system and how its components interact. Figure 10-5. A production-grade agentic system Let’s explore each of these types. Models Language models are the backbone of agentic systems, responsible for their autonomous nature and problem-solving capabilities. A single agentic system could be composed of multiple language models, with each model playing a distinct role. For example, you can build an