#Context Windows

當 RAG 不夠用時：快取增強生成（CAG）、混合檢索與工作記憶

基本的檢索增強生成（RAG）到現在還是多數正式環境系統最常用的基礎模式。語料庫很大、更新頻繁、又需要展示答案來源？檢索依然是最乾淨的起點。不過實務上會碰到一種極限情境——簡單的 RAG 開始力不從心：系統確實找到了正確的文件，但任務現在需要的是針對一組有限的證據持續推理，同時回答同一案件的多次後續追問。

記憶、狀態與知識：別再把所有東西都叫做「記憶」

一個為中型旅行社打造的旅遊規劃 Copilot，被問了一個很直接的問題：「這間飯店之前是否在輪椅使用者的無障礙審查中未通過？」

提示詞、上下文窗口，以及你如何與 LLM 對話

上一篇裡，我們丟了一句話給 LLM——「幫我規劃一趟赫爾辛基(Helsinki)之旅」——然後拿到一份細節滿滿的行程表：餐廳名、交通路線、一日遊安排。讀起來很順，看起來也合理，但好幾個細節事後被證實是錯的。模型沒壞，只是輸入沒給它什麼限制條件可以依循。

When RAG Is Not Enough: CAG, Hybrid Retrieval, and Working Memory

Basic retrieval-augmented generation, or RAG, is still the default grounding pattern for most production systems. If you have a large corpus, frequent updates, and a need to show where an answer came from, retrieval remains the cleanest starting point. But there is a practical limit case where si...

Memory, State, and Knowledge: Stop Calling Everything "Memory"

A travel-planning copilot for a mid-size agency is asked a straightforward question: "Did this hotel fail an accessibility review before for wheelchair users?"

Prompts, Context Windows, and How You Talk to an LLM

In the previous post, we sent a single line to an LLM — "Plan a trip to Helsinki" — and got back an itinerary full of specific-sounding details: restaurant names, transit directions, day-trip logistics. It was fluent and plausible, but several of those details turned out to be wrong. The model wa...