大上下文窗口的"智能盲区"陷阱
推荐指数 84.0 NO. 013 · 2026.06.15
发布2026/06/14Score230Comments173
为什么值得看
LLM上下文窗口存在"智能区"（约10万token内）和"遗忘区"的隐性分界，编程Agent极易因长会话滑入后者。这直接威胁依赖长上下文的代码生成、多轮调试等Agent工作流的可靠性。
编辑判断
这个问题被行业长期忽视是因为基准测试的欺骗性：needle-in-haystack测试只测单点检索，不测多轮推理中的注意力衰减。实际工程中，Cursor、Windsurf等Agent的"自动读取相关文件"机制正是陷阱——用户以为模型"看到了"全部代码，实则在长会话后已丢失早期文件细节。
更隐蔽的风险在于成本结构：Claude 3.5 Sonnet的200K窗口让团队倾向于把所有代码塞进去，而非做RAG或代码切片，结果是用3倍token换更差的准确性。建议给Agent会话硬编码token上限告警（如80K），强制触发上下文压缩或重启会话。
社区反馈
意见分歧 155 条评论
核心争论：上下文"智能区"边界究竟在哪——是10万、60万还是近百万token，模型版本差异巨大
Perhaps compacting the context can be made in multiple requests over smaller and overlapping chunks to avoid using the 'dumb zone', and for yielding a better result.
I /clear all the time out of habit. I want to be able to get the thing done with minimal context. It also means you can do it again slightly different if needed, you know the seed conditions for the task.
The approach we're taking to deal with this very real context rot is using a bunch of related techniques which we call transposing the agent loop: https://alejo.ch/3jt In essence, we run many short agent loops, generating their prompts dynamically from structured data. Each loop advan