新增wiki命令文件
This commit is contained in:
42
wiki/concepts/Context-Window.md
Normal file
42
wiki/concepts/Context-Window.md
Normal file
@@ -0,0 +1,42 @@
|
||||
---
|
||||
title: "Context Window"
|
||||
type: concept
|
||||
tags: [AI, 模型, Context]
|
||||
sources: [yang-xia-ri-ji-4, yang-long-xia-5-tian]
|
||||
last_updated: 2026-04-14
|
||||
---
|
||||
|
||||
## Definition
|
||||
模型能处理的上下文大小,决定了对话长度和信息保留量。
|
||||
|
||||
## 典型值
|
||||
| 模型 | Context Window |
|
||||
|------|---------------|
|
||||
| MiniMax-M2.7 | 200K tokens |
|
||||
| deepseek-reasoner | 16K tokens |
|
||||
|
||||
## Compaction机制
|
||||
OpenClaw的对话压缩机制:
|
||||
- 当对话填满Context Window时,将旧消息压缩成摘要
|
||||
- 摘要抓住要点,但丢失细节(姓名、数字、具体决定)
|
||||
- safeguard模式预留一半token给compaction
|
||||
|
||||
## Memory Flush解决方案
|
||||
压缩前将重要上下文写入磁盘:
|
||||
```json
|
||||
{
|
||||
"compaction": {
|
||||
"memoryFlush": {
|
||||
"enabled": true,
|
||||
"softThresholdTokens": 4000
|
||||
}
|
||||
}
|
||||
}
|
||||
```
|
||||
|
||||
## Connections
|
||||
- [[Compaction]] ← 上下文压缩 ← [[Context-Window]]
|
||||
- [[Memory-Flush]] ← 保护机制 ← [[Context-Window]]
|
||||
- [[Embedding Vector]] ← 限制 ← [[Context-Window]]
|
||||
- [[Generation]] ← 限制 ← [[Context-Window]]
|
||||
- [[Indexing]] ← 限制 ← [[Context-Window]]
|
||||
Reference in New Issue
Block a user