Files
myWiki/concepts/context-prefetch-vs-agentic.md

39 lines
1.5 KiB
Markdown
Raw Blame History

This file contains ambiguous Unicode characters

This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

---
title: "上下文预取 vs 按需加载Context Prefetch vs Agentic"
created: 2026-07-03
updated: 2026-07-03
type: concept
tags: ["agent", "context", "latency", "token-economics"]
sources: ["https://mp.weixin.qq.com/s/iiTmgbtrYHMMjQ7dn7CDrg"]
---
# 上下文预取 vs 按需加载
> Harness 要明确规定哪些提前带入Prefetch哪些按需读取Agentic
## 核心定义
在 [[workspace-first-architecture|Workspace-first 架构]] 中,上下文加载被拆成两种模式:
| 模式 | 触发时机 | 内容 | 特点 |
|------|---------|------|------|
| **Prefetch** | 进入工作区时自动 | 用户偏好、近期工作事件、常用经验 | 短、准、可控 |
| **Agentic** | 模型主动请求 | 完整文档详情、历史会议纪要 | 按需触发,增加交互轮次 |
## 设计权衡
- 预取过多 → 抬高上下文成本,大量不相关内容占据注意力
- 全部按需读取 → 增加交互轮次和失败点,延迟上升
- 最优策略:核心信息 prefetch + 详情按需 agentic
## 与记忆召回的关系
Prefetch/Agentic 是上下文层的加载策略,[[memory-recall-fast-slow|记忆快慢召回]] 是记忆层对同一思路的实现——fast 对应 prefetchslow 对应 agentic。
## 参考
- [[zleap-workspace-harness-2026|Zleap-Agent Harness 设计]]
- [[workspace-first-architecture|Workspace-first 架构]]
- [[context-management|上下文管理]]
- [[memory-recall-fast-slow|快慢召回]]