Ci bi mu melGis bi ci topp
Kontekst bu gudd vs RAG
IA làkk
Làkk AI GUIDE
Context engineering mooy xelal bi ñuy def ngir xam li modelu bopp bi gis ci yoon bu nekk (tënk-tënk, yoon yi ñu jël, definition ak natu tool yi, xalaat ak xam-xamu bopp bi) te dëgër ko ci window bu amul benn. Mooy li gën a am solo ci li model bi gis.
Déglul tegtal biiIA-nettali · ~3 min
Mooy li gën a am solo ci li model bi gis. Mooy li gën a am solo ndax output quality bi dafa aju ci li nekk ci window bi, te ci agents ak long tasks, context bi ñu defar ko ak code ci call bu nekk, du nit ki def ko benn yoon.
Prompt engineering focuses on wording a single request. Context engineering takes a wider view. The model receives one sequence of tokens, and everything in it competes for space and for the model's attention. That sequence usually includes a system prompt, tool definitions, retrieved passages, prior turns, tool outputs and sometimes long-term memory. The term became common in 2025, when teams building agents found that many failures came from the model having the wrong information, too much of it or conflicting information, rather than from poor phrasing. Several principles follow. First, relevance beats volume. Some models accept hundreds of thousands of tokens, but they do not use every token equally well. The 'lost in the middle' study (Liu et al., 2023) found that models often use information at the start and end of a long input better than information buried in the middle. Second, context is a budget. Every token costs money and adds latency, and irrelevant material can distract the model. Third, structure helps. Clear sections, labeled documents and consistent formats make it easier for the model to find what matters. Common techniques include: - retrieval-augmented generation, which fetches only the relevant passages - just-in-time loading, where an agent holds lightweight references such as file names or IDs and uses tools to pull full content when needed - structured note-taking that keeps state outside the window - compaction of long histories - sub-agents that explore in their own context and return condensed findings One misconception is that context engineering is just a new name for prompting. Wording is one part of it, but most of the work is systems design: deciding what gets retrieved, summarized or dropped, and in what order it appears. Another misconception is that a bigger window removes the need for this. A bigger window lowers the pressure, but cost and attention limits remain.
Bot bu joxe doole ci client yi day jël rekk ñetti yoon yi gën a am solo ci help-center bi ak plan tier bu client bi. Du defar knowledge base bi yépp ci request bu nekk.
Agent bu code day tambali ak liste yoon yi ak function signatures, ba noppi day jëfandikoo read-file tool ngir jël file yi yépp rekk bu ñu am solo.
Assistant bu meeting day denc decision ak action items ci file bu am structure, ba noppi day jël file bi ci session bi ci topp, du defar transcript yi yépp.
Team bi day soppi database tool bi ngir mu jël filtered summary ak row count, du 5,000 rows bu raw JSON. Lii day tax window bu agent bi du dëgër ak data yu du jëfandikoo.
Context windows yi dañu gën a mag, te models yi dañu gën a am solo ci jëfandikoo long inputs. Cost, latency ak attention limits yi dañuy tax mu am solo ngir tànn li ñuy def. Gën a bari ci xelal bi ñuy def ci frameworks ak platforms, te bari providers yi ñu am memory tools, automatic compaction ak managed retrieval. Xelal bi day soppi ci defar context bi ci anam gu amul benn, ba noppi day defar policies ngir xam li agent bi wara xam, jël walla wàcc, ba noppi day test policies yi. Amul benn li ñu xam ndax future model designs yi ñu amul benn. Ci anam gu nekk, context bi ñu defar ko ci anam gu amul benn mooy li gën a am solo ngir gën a am solo ci output bu model bi.
Bulletin bu amul fayda
Ñatti jaar-jaari IA yuñ firndeel bis bu nekk ci ayu-bis bi, ñu bind leen ci Àngle bu leer. Gratis ba fàww, amul yëgle.
Benn email bis bu nekk ci ayu-bis bi. Benn klik nga dindi abonemaa bi. Dunu musa jaay wala séddoo sa adres.
Nattal sa bopp
Feedback ci saasi ci bépp tontu, ak sertifikaa buñ mëna séddoo ak ID buñ mëna saytu soo jàllee ci njàng mi.
Jàppale njàngum IA te doo fay. AI Understanding mbootaay la buy def te Yàlla tax — amul yëgle, amul paywall, musul am. Defal ab sarax
Context engineering mooy xelal bi ñuy def ngir xam li modelu bopp bi gis ci yoon bu nekk (tënk-tënk, yoon yi ñu jël, definition ak natu tool yi, xalaat ak xam-xamu bopp bi) te dëgër ko ci window bu amul benn. Mooy li gën a am solo ci li model bi gis. Mooy li gën a am solo ndax output quality bi dafa aju ci li nekk ci window bi, te ci agents ak long tasks, context bi ñu defar ko ak code ci call bu nekk, du nit ki def ko benn yoon.
Context engineering dafay jàppee xeetu token bi yépp model bi jot ni luñu wara jëmmal: tegtal, passage yuñ jëlee, joxe leeral ci jumtukaay ak ay génne, jaar-jaar ak fàttaliku. Baatu-kàddu yu gaaw yi benn wàll la ci.
Liu ak ñeneen. (2023) dafa wane ni performance dafay faral di wàññeeku sudee xibaar bi am solo mingi toog ci diggu muy lu yàgg. Loolu mooy benn ci sabab yi pertinence ak plasement gëna am solo ci volume bi.
Jeton yi ci palanteer bi amnañu njëg dëgg ci njëg ak diiru tontu. Jetons yu amul njariñ itam dañuy joŋante ak yi am njariñ ngir model bi bàyyi seen xel ci, kon dañu leen wara jëfandikoo ci jom.
Du jël yépp ci tambali, agent bi day am pointers, ba noppi day jël full content rekk bu step bi am solo. Lii day tax window bi du dëgër.
Providers yi ñu am prompt caching ñu am prefix bu nekk byte-identical ci call yi. Def stable content ci tambali, conversation bi ci topp, day tax mu am solo.
Weyal di jàng
Tann nañu yeneen njiit ngir topic bii
Ci bi mu melGis bi ci topp
Kontekst bu gudd vs RAG
IA làkk