[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"portal-settings:stajic:en":3,"public-menus:all":38,"post:rag-failed-but-which-layer-actually-failed-a-diagnostic-method:en":205,"related:post:rag-failed-but-which-layer-actually-failed-a-diagnostic-method:en:1":1171},{"statusCode":4,"data":5,"message":37},200,{"tenantId":6,"lang":7,"defaultLang":8,"siteUrl":9,"contactEmail":10,"brandName":11,"logoUrl":12,"siteName":11,"siteDescription":13,"ogImage":10,"robotsIndex":14,"socialLinks":10,"reservedSlugs":10,"seoPolicy":15},"stajic","en","de","https:\u002F\u002Fstajic.de",null,"Stajic Platform","\u002FLogo_Planet.svg","Stajic Portal",true,{"branding":16,"relatedContent":17,"crossDomainLinks":18},{"logoUrl":12},{"enabled":14},[19,22,25,28,31,34],{"url":20,"label":21,"isActive":14,"showInFooter":14,"includeInSameAs":14},"https:\u002F\u002Ffigure.rocks","figure.rocks",{"url":23,"label":24,"isActive":14,"showInFooter":14,"includeInSameAs":14},"https:\u002F\u002Floving.rocks","loving.rocks",{"url":26,"label":27,"isActive":14,"showInFooter":14,"includeInSameAs":14},"https:\u002F\u002Fbazify.com","bazify.com",{"url":29,"label":30,"isActive":14,"showInFooter":14,"includeInSameAs":14},"https:\u002F\u002Fbazify.de","bazify.de",{"url":32,"label":33,"isActive":14,"showInFooter":14,"includeInSameAs":14},"https:\u002F\u002Fbazify.at","bazify.at",{"url":35,"label":36,"isActive":14,"showInFooter":14,"includeInSameAs":14},"https:\u002F\u002Fbazify.ba","bazify.ba","Portal settings resolved",[39,45],{"id":40,"name":41,"location":42,"isActive":14,"isDefault":43,"items":44},1,"main-navigation","header",false,[],{"id":46,"name":47,"location":48,"isActive":14,"isDefault":14,"items":49},4,"main-menu","sidebar",[50,66,79,93,103,118,133],{"id":51,"title":52,"url":60,"target":61,"icon":62,"isActive":14,"type":63,"productId":10,"categoryId":10,"shopCategoryId":10,"articleId":10,"pageId":64,"portfolioId":10,"children":65},"item-18",{"de":53,"en":54,"es":55,"fr":56,"it":54,"ru":57,"sr":58,"zh":59},"Startseite","Home","Inicio","Accueil","Главная","Почетна","首页","\u002Ffull-stack-web-developer-munich-performance-seo-and-maintainable-builds","_self","i-lucide-home","page",111,[],{"id":67,"title":68,"url":75,"target":61,"icon":76,"isActive":14,"type":63,"productId":10,"categoryId":10,"shopCategoryId":10,"articleId":10,"pageId":77,"portfolioId":10,"children":78},"item-22",{"de":69,"en":69,"es":70,"fr":69,"it":71,"ru":72,"sr":73,"zh":74},"Vision","Visión","Visione","Видение","Визија","想象","\u002Fueber-uns-webdesign-muenchen-webaplikation","i-lucide-eye",113,[],{"id":80,"title":81,"url":89,"target":61,"icon":90,"isActive":14,"type":63,"productId":10,"categoryId":10,"shopCategoryId":10,"articleId":10,"pageId":91,"portfolioId":10,"children":92},"item-19",{"de":82,"en":83,"es":84,"fr":83,"it":85,"ru":86,"sr":87,"zh":88},"Leistungen","Services","Servicios","Servizi","Услуги","Услуге","服务","\u002Fservices-dienstleistungen-muenchen","i-lucide-wrench",116,[],{"id":94,"title":95,"url":99,"target":61,"icon":100,"isActive":14,"type":63,"productId":10,"categoryId":10,"shopCategoryId":10,"articleId":10,"pageId":101,"portfolioId":10,"children":102},"item-23",{"de":96,"en":96,"es":96,"fr":96,"it":96,"ru":97,"sr":97,"zh":98},"Blog","Блог","博客","\u002Fblog","i-lucide-book-open",112,[],{"id":104,"title":105,"url":114,"target":61,"icon":115,"isActive":14,"type":63,"productId":10,"categoryId":10,"shopCategoryId":10,"articleId":10,"pageId":116,"portfolioId":10,"children":117},"item-32",{"de":106,"en":107,"es":108,"fr":109,"it":110,"ru":111,"sr":112,"zh":113},"Neue Technologien","New Technologies","Nuevas tecnologías","Nouvelles technologies","Nuove tecnologie","Новые технологии","Нове технологије","新技术！","\u002Fneue-webtechnologien","i-lucide-sparkles",122,[],{"id":119,"title":120,"url":129,"target":61,"icon":130,"isActive":14,"type":63,"productId":10,"categoryId":10,"shopCategoryId":10,"articleId":10,"pageId":131,"portfolioId":10,"children":132},"item-20",{"de":121,"en":122,"es":123,"fr":124,"it":125,"ru":126,"sr":127,"zh":128},"Kontakt","Contact us!","Contacto","Contact","Contatto","Контакт","Контактирајте нас","联系我们！","\u002Fcontact","i-lucide-mail",115,[],{"id":134,"title":135,"url":144,"target":61,"icon":145,"isActive":14,"type":63,"productId":10,"categoryId":10,"shopCategoryId":10,"articleId":10,"pageId":146,"portfolioId":10,"children":147},"item-21",{"de":136,"en":137,"es":138,"fr":139,"it":140,"ru":141,"sr":142,"zh":143},"Unsere Arbeit","Our Work","Nuestro trabajo","Nos réalisations","I nostri lavori","Наши работы","Наши радови","文件夹","\u002Fportfolio","i-lucide-briefcase",114,[148,161,175,181,193],{"id":149,"title":150,"url":144,"target":61,"icon":159,"isActive":14,"type":63,"productId":10,"categoryId":10,"shopCategoryId":10,"articleId":10,"pageId":146,"portfolioId":10,"children":160},"item-24",{"de":151,"en":152,"es":153,"fr":154,"it":155,"ru":156,"sr":157,"zh":158},"Alle Projekte","All Projects","Todos los proyectos","Tous les projets","Tutti i progetti","Все проекты","Сви пројекти","所有项目","i-lucide-grid-3x3",[],{"id":162,"title":163,"url":171,"target":61,"icon":172,"isActive":14,"type":173,"productId":10,"categoryId":10,"shopCategoryId":10,"articleId":10,"pageId":10,"portfolioId":10,"children":174},"item-29",{"de":164,"en":165,"es":166,"fr":167,"it":168,"ru":169,"sr":170,"zh":143},"Local Roots, Global Reach","Local Roots - Global Reach","Empresa local ","Entreprise locale","Azienda locale","Местная компания","Локално предузеће глобално тржиште","\u002Fportfolio\u002Flocal-roots-global-reach-communication-media-systems-for-modern-business","i-lucide-folder","custom",[],{"id":176,"title":177,"url":179,"target":61,"icon":172,"isActive":14,"type":173,"productId":10,"categoryId":10,"shopCategoryId":10,"articleId":10,"pageId":10,"portfolioId":10,"children":180},"item-28",{"de":178,"en":178,"es":178,"fr":178,"it":178,"ru":178,"sr":178,"zh":178},"Solr Suggester","\u002Fportfolio\u002Fsolr-fuzzy-suggester-und-solr-infix-suggester-abfrage-ueber-ajax-und-filterung",[],{"id":182,"title":183,"url":191,"target":61,"icon":172,"isActive":14,"type":173,"productId":10,"categoryId":10,"shopCategoryId":10,"articleId":10,"pageId":10,"portfolioId":10,"children":192},"item-27",{"de":184,"en":185,"es":186,"fr":187,"it":188,"ru":189,"sr":190,"zh":185},"Firmenwebseite SEO","Company Website SEO","Sitio web corporativo SEO","Site web d’entreprise SEO","Sito web aziendale SEO","Корпоративный сайт SEO","Пословна веб-страница SEO","\u002Fportfolio\u002Fseo-sem-branding-mobile-webseite-muenchen",[],{"id":194,"title":195,"url":203,"target":61,"icon":172,"isActive":14,"type":173,"productId":10,"categoryId":10,"shopCategoryId":10,"articleId":10,"pageId":10,"portfolioId":10,"children":204},"item-31",{"de":196,"en":197,"es":198,"fr":199,"it":200,"ru":201,"sr":202,"zh":197},"Digitalisierungsportal","Digitalization Portal","Portal de digitalización","Portail de numérisation","Portale di digitalizzazione","Портал цифровизации","Портал за дигитализацију","\u002Fportfolio\u002Fdigitalisierungsportal-archiv-museum-bibliothek-ead-lido-mets-mods",[],{"statusCode":4,"data":206,"message":1170},{"id":207,"title":208,"slug":209,"content":210,"contentJson":211,"excerpt":834,"featuredImage":835,"featuredImageAlt":836,"featuredImageCaption":10,"featuredImageTitle":10,"featuredImageCopyright":10,"featuredImageAuthor":10,"featuredImageSourceUrl":10,"featuredImageLicense":10,"featuredImageIsAiGenerated":43,"status":837,"publishedAt":838,"createdAt":839,"updatedAt":840,"seoLocalePaths":841,"categories":850,"author":863,"translations":868},"469","RAG Failed — But Which Layer Actually Failed? A Diagnostic Method","rag-failed-but-which-layer-actually-failed-a-diagnostic-method","{\"time\":1790369097340,\"blocks\":[{\"id\":\"8zyFXn5HD5\",\"type\":\"tableOfContents\",\"data\":{\"title\":\"Contents\",\"minLevel\":2,\"maxLevel\":3},\"tunes\":{}},{\"id\":\"intro\",\"type\":\"paragraph\",\"data\":{\"text\":\"A RAG system returns a weak, wrong, incomplete, or unsupported answer. The usual diagnosis is “retrieval failed” or “the model hallucinated.” Both labels are too broad to be useful. A production RAG pipeline can fail before retrieval, during retrieval, while ranking, while assembling context, during generation, or after generation when evidence and validity are checked.\"},\"tunes\":{}},{\"id\":\"direct\",\"type\":\"callout\",\"data\":{\"variant\":\"info\",\"title\":\"Direct answer\",\"body\":\"\u003Cstrong>Do not debug RAG as one component.\u003C\u002Fstrong> Diagnose it as a chain of independently testable layers. First determine whether the required evidence exists in an authoritative source. Then test query construction, candidate retrieval, ranking, context assembly, generation, evidence attribution, and freshness. The fastest isolation technique is an \u003Cstrong>oracle-context test\u003C\u002Fstrong>: give the generator the correct evidence manually. If the answer becomes correct, the dominant failure is upstream of generation. If it remains wrong, retrieval is not the primary problem.\"},\"tunes\":{}},{\"id\":\"model-note\",\"type\":\"callout\",\"data\":{\"variant\":\"note\",\"title\":\"About the diagnostic model\",\"body\":\"The RAG Failure Stack in this article is a practical diagnostic model, not a formal industry standard. Existing platforms already separate retrieval-only metrics from retrieve-and-generate metrics; this model extends that separation into a step-by-step production debugging method.\"},\"tunes\":{}},{\"id\":\"h-not-diagnosis\",\"type\":\"header\",\"data\":{\"text\":\"Why “RAG failed” is not a diagnosis\",\"level\":2},\"tunes\":{}},{\"id\":\"p-not-1\",\"type\":\"paragraph\",\"data\":{\"text\":\"Retrieval-augmented generation combines several mechanisms: a user request is interpreted, one or more searches are constructed, candidate material is retrieved, results are filtered or reranked, selected evidence is inserted into a model context, and a model generates an answer. Production systems may add permissions, metadata filters, freshness rules, citations, query rewriting, hybrid search, tool calls, memory, and external state.\"},\"tunes\":{}},{\"id\":\"p-not-2\",\"type\":\"paragraph\",\"data\":{\"text\":\"A wrong final answer therefore does not tell you which component failed. The model may have received the wrong evidence. It may have received the right evidence mixed with too much noise. The evidence may be correct but stale. The source may never have contained the answer. Or the model may have ignored perfectly adequate context.\"},\"tunes\":{}},{\"id\":\"p-not-3\",\"type\":\"paragraph\",\"data\":{\"text\":\"OpenAI's RAG guidance already makes a fundamental distinction between retrieval failure and model failure: a system can supply the wrong context, or it can supply the right context and still generate the wrong answer. AWS similarly separates retrieve-only evaluation from retrieve-and-generate evaluation. For production diagnosis, that distinction should be taken further.\"},\"tunes\":{}},{\"id\":\"h-stack\",\"type\":\"header\",\"data\":{\"text\":\"The RAG Failure Stack\",\"level\":2},\"tunes\":{}},{\"id\":\"table-stack\",\"type\":\"table\",\"data\":{\"withHeadings\":true,\"stretched\":false,\"content\":[[\"Layer\",\"Question\",\"Typical failure\"],[\"1. Source coverage\",\"Does the required evidence exist in an allowed authoritative source?\",\"The corpus cannot answer the question at all\"],[\"2. Query construction\",\"Did the system search for the right thing?\",\"Intent, entities, filters, language, or time constraints are lost\"],[\"3. Candidate retrieval\",\"Did the relevant evidence enter the candidate set?\",\"Low recall; the right chunk is never retrieved\"],[\"4. Ranking & filtering\",\"Did the right evidence survive and rank high enough?\",\"Relevant evidence is buried, filtered out, or outranked by superficially similar text\"],[\"5. Context assembly\",\"Did the model receive usable evidence?\",\"Truncation, bad chunk boundaries, duplicates, conflicting passages, or context overload\"],[\"6. Generation\",\"Did the model use the supplied evidence correctly?\",\"Unsupported inference, instruction failure, reasoning error, or refusal mismatch\"],[\"7. Evidence attribution\",\"Can the answer be traced to the evidence it claims to use?\",\"Missing, weak, or incorrect citations; claims exceed retrieved support\"],[\"8. Validity & freshness\",\"Is the evidence still valid for this question now?\",\"Correct historical evidence is reused outside its valid time, version, jurisdiction, or state\"]]},\"tunes\":{}},{\"id\":\"h-source\",\"type\":\"header\",\"data\":{\"text\":\"Layer 1 — Source coverage: can the system answer this at all?\",\"level\":3},\"tunes\":{}},{\"id\":\"p-source-1\",\"type\":\"paragraph\",\"data\":{\"text\":\"Before tuning embeddings, rerankers, or prompts, verify that the answer exists in the knowledge space the system is allowed to use. This sounds obvious, but many RAG failures are actually corpus failures. The requested fact may be absent, hidden in an unindexed attachment, available only in a newer document, stored in a system outside the RAG corpus, or blocked by permissions.\"},\"tunes\":{}},{\"id\":\"p-source-2\",\"type\":\"paragraph\",\"data\":{\"text\":\"A retrieval metric cannot recover information that was never indexed. A larger top-k cannot retrieve a document the pipeline does not contain. If the source coverage test fails, the correct fix is ingestion, source selection, permissions, or an explicit “not answerable from available evidence” behaviour.\"},\"tunes\":{}},{\"id\":\"source-warning\",\"type\":\"callout\",\"data\":{\"variant\":\"warning\",\"title\":\"Failure pattern\",\"body\":\"Teams often tune retrieval against questions that the corpus cannot actually answer. This can make the retriever better at finding related text while leaving the underlying information gap untouched.\"},\"tunes\":{}},{\"id\":\"h-query\",\"type\":\"header\",\"data\":{\"text\":\"Layer 2 — Query construction: did the system ask the corpus the right question?\",\"level\":3},\"tunes\":{}},{\"id\":\"p-query-1\",\"type\":\"paragraph\",\"data\":{\"text\":\"The user query is not always the retrieval query. Production systems rewrite questions, resolve pronouns, extract entities, translate languages, add metadata constraints, split complex questions, or generate multiple searches. Every transformation can improve retrieval, but every transformation can also destroy information.\"},\"tunes\":{}},{\"id\":\"p-query-2\",\"type\":\"paragraph\",\"data\":{\"text\":\"A request such as “Does the policy still apply to contractors in Germany after the September update?” contains at least an entity, a population, a jurisdiction, and a time boundary. A rewritten query that becomes “contractor policy” may retrieve semantically related text while losing the variables that decide whether the answer is valid.\"},\"tunes\":{}},{\"id\":\"h-retrieval\",\"type\":\"header\",\"data\":{\"text\":\"Layer 3 — Candidate retrieval: did the relevant evidence enter the set?\",\"level\":3},\"tunes\":{}},{\"id\":\"p-ret-1\",\"type\":\"paragraph\",\"data\":{\"text\":\"Candidate retrieval is primarily a recall problem. The diagnostic question is not yet whether the best result ranked first; it is whether relevant evidence appeared anywhere in the candidate pool. If the known correct source does not appear, investigate indexing, chunking, embeddings, lexical matching, metadata, hybrid search, language handling, synonyms, and query expansion.\"},\"tunes\":{}},{\"id\":\"p-ret-2\",\"type\":\"paragraph\",\"data\":{\"text\":\"This is where retrieval-only evaluation is valuable. AWS exposes context relevance and context coverage for retrieve-only RAG evaluation. The important production habit is to evaluate retrieval before generation so that a polished final answer cannot hide a weak candidate set.\"},\"tunes\":{}},{\"id\":\"h-ranking\",\"type\":\"header\",\"data\":{\"text\":\"Layer 4 — Ranking and filtering: was the right evidence discarded or buried?\",\"level\":3},\"tunes\":{}},{\"id\":\"p-rank-1\",\"type\":\"paragraph\",\"data\":{\"text\":\"A system can have good recall and still fail because the relevant evidence ranks below noisy but semantically similar material. Rerankers, recency boosts, authority weights, language preferences, tenant filters, access controls, product status filters, and deduplication all change what survives into the final context.\"},\"tunes\":{}},{\"id\":\"p-rank-2\",\"type\":\"paragraph\",\"data\":{\"text\":\"Debugging should therefore preserve the full candidate list, not only the final top-k. If the gold evidence was retrieved at rank 18 and a reranker removed it, the fix is not the same as a retrieval miss.\"},\"tunes\":{}},{\"id\":\"h-context\",\"type\":\"header\",\"data\":{\"text\":\"Layer 5 — Context assembly: did useful evidence become usable context?\",\"level\":3},\"tunes\":{}},{\"id\":\"p-ctx-1\",\"type\":\"paragraph\",\"data\":{\"text\":\"Retrieval success does not guarantee context success. Relevant chunks can be truncated, separated from their qualifiers, duplicated until they dominate the prompt, mixed with contradictory versions, or surrounded by enough irrelevant text that the decisive passage loses salience.\"},\"tunes\":{}},{\"id\":\"p-ctx-2\",\"type\":\"paragraph\",\"data\":{\"text\":\"Chunk boundaries are especially important. A sentence may contain the rule while the following sentence contains the exception. If they are indexed separately and only the first is retrieved, the retriever can appear relevant while the assembled context becomes misleading.\"},\"tunes\":{}},{\"id\":\"h-generation\",\"type\":\"header\",\"data\":{\"text\":\"Layer 6 — Generation: can the model use correct evidence correctly?\",\"level\":3},\"tunes\":{}},{\"id\":\"p-gen-1\",\"type\":\"paragraph\",\"data\":{\"text\":\"Once the system has demonstrably supplied sufficient evidence, generation becomes independently testable. The model may overgeneralize, combine incompatible passages, ignore a negative statement, fail to follow the requested answer format, invent a bridge between facts, or answer from parametric memory instead of the retrieved evidence.\"},\"tunes\":{}},{\"id\":\"p-gen-2\",\"type\":\"paragraph\",\"data\":{\"text\":\"This is why end-to-end correctness alone is insufficient for diagnosis. OpenAI recommends evaluation as a structured way to understand application behaviour, while Anthropic's agent-evaluation guidance emphasizes multiple trials, graders, traces, and realistic failure cases. For RAG, the generator should be tested both with normal retrieval and with controlled gold context.\"},\"tunes\":{}},{\"id\":\"h-evidence\",\"type\":\"header\",\"data\":{\"text\":\"Layer 7 — Evidence attribution: is the answer actually supported?\",\"level\":3},\"tunes\":{}},{\"id\":\"p-evidence-1\",\"type\":\"paragraph\",\"data\":{\"text\":\"A plausible answer with citations can still be weakly grounded. The cited document may be relevant to the topic but not support the specific claim. One sentence may be supported while another is inferred. A citation may point to a source that contradicts the answer once its conditions are read.\"},\"tunes\":{}},{\"id\":\"p-evidence-2\",\"type\":\"paragraph\",\"data\":{\"text\":\"Citation evaluation therefore belongs after generation. AWS distinguishes citation precision from citation coverage: whether cited passages are correctly cited and whether the answer is sufficiently supported by citations. In production, claim-level support is more useful than treating the presence of any citation as evidence quality.\"},\"tunes\":{}},{\"id\":\"h-validity\",\"type\":\"header\",\"data\":{\"text\":\"Layer 8 — Validity and freshness: was the evidence correct for this version of reality?\",\"level\":3},\"tunes\":{}},{\"id\":\"p-valid-1\",\"type\":\"paragraph\",\"data\":{\"text\":\"RAG can retrieve a perfectly authentic, highly relevant, faithfully quoted source and still produce a wrong answer if the source is no longer valid for the current question. Policies change. APIs are deprecated. prices move. software behaviour changes between versions. product inventory changes. permissions change. game patches change mechanics.\"},\"tunes\":{}},{\"id\":\"p-valid-2\",\"type\":\"paragraph\",\"data\":{\"text\":\"This is a separate failure class from hallucination. The evidence is real; its applicability is wrong. A robust system therefore needs timestamps, version or jurisdiction metadata where relevant, source authority, supersession rules, and an explicit mechanism for deciding when older evidence must be restricted or abandoned.\"},\"tunes\":{}},{\"id\":\"h-oracle\",\"type\":\"header\",\"data\":{\"text\":\"The fastest isolation method: the oracle-context test\",\"level\":2},\"tunes\":{}},{\"id\":\"p-oracle-1\",\"type\":\"paragraph\",\"data\":{\"text\":\"The most useful first split is simple: manually provide the generator with a small set of evidence that you know is sufficient to answer the question. Keep the task and expected answer unchanged.\"},\"tunes\":{}},{\"id\":\"oracle-comparison\",\"type\":\"comparison\",\"data\":{\"title\":\"Oracle-context test\",\"layout\":\"table\",\"columns\":[{\"id\":\"result\",\"label\":\"Result\"},{\"id\":\"meaning\",\"label\":\"Likely interpretation\"},{\"id\":\"next\",\"label\":\"Next diagnostic step\"}],\"rows\":[{\"id\":\"oracle-pass\",\"label\":\"Answer becomes correct\",\"values\":[\"\",\"\",\"\"]},{\"id\":\"oracle-fail\",\"label\":\"Answer remains wrong\",\"values\":[\"\",\"\",\"\"]},{\"id\":\"oracle-partial\",\"label\":\"Answer improves but remains incomplete\",\"values\":[\"\",\"\",\"\"]}]},\"tunes\":{}},{\"id\":\"oracle-tip\",\"type\":\"callout\",\"data\":{\"variant\":\"tip\",\"title\":\"Why this test is powerful\",\"body\":\"The oracle-context test removes most of the retrieval pipeline from the experiment. It does not prove that generation is perfect, but it gives you a fast counterfactual: \u003Cstrong>what would the model do if retrieval had already succeeded?\u003C\u002Fstrong>\"},\"tunes\":{}},{\"id\":\"h-sequence\",\"type\":\"header\",\"data\":{\"text\":\"A production diagnostic sequence\",\"level\":2},\"tunes\":{}},{\"id\":\"diag-flow\",\"type\":\"processFlow\",\"data\":{\"title\":\"Diagnose the failure from evidence to answer\",\"orientation\":\"auto\",\"steps\":[{\"label\":\"1. Define the expected claim\",\"description\":\"Write the expected answer, allowed uncertainty, and the evidence that would justify it.\"},{\"label\":\"2. Verify source coverage\",\"description\":\"Confirm that authoritative and permitted evidence exists in the indexed or reachable source set.\"},{\"label\":\"3. Run the oracle-context test\",\"description\":\"Supply sufficient gold evidence directly to the generator and observe whether the answer becomes correct.\"},{\"label\":\"4. Inspect the retrieval query\",\"description\":\"Check rewrites, entities, filters, language, time constraints, decomposition, and hidden assumptions.\"},{\"label\":\"5. Inspect candidates before reranking\",\"description\":\"Determine whether relevant evidence was retrieved at all and record its rank.\"},{\"label\":\"6. Inspect ranking and context assembly\",\"description\":\"Check reranking, metadata filters, truncation, chunk boundaries, duplicates, conflicts, and top-k composition.\"},{\"label\":\"7. Grade generation and citations separately\",\"description\":\"Measure answer correctness, completeness, faithfulness, and claim-level evidence support.\"},{\"label\":\"8. Test validity boundaries\",\"description\":\"Check whether version, date, state, jurisdiction, permissions, or superseding evidence changes the answer.\"}]},\"tunes\":{}},{\"id\":\"h-one-change\",\"type\":\"header\",\"data\":{\"text\":\"Do not change three layers at once\",\"level\":2},\"tunes\":{}},{\"id\":\"p-one-1\",\"type\":\"paragraph\",\"data\":{\"text\":\"A common debugging mistake is to change embeddings, chunk sizes, top-k, prompts, and the model in one iteration. If the score improves, you do not know why. If it gets worse, you do not know which change caused the regression.\"},\"tunes\":{}},{\"id\":\"p-one-2\",\"type\":\"paragraph\",\"data\":{\"text\":\"Treat RAG debugging like experimental diagnosis: hold as much of the pipeline constant as possible and replace one uncertain component with a controlled input. Gold documents isolate retrieval. Gold chunks isolate chunk selection. Fixed context isolates generation. A fixed model isolates retrieval changes. A fixed corpus isolates ingestion and indexing changes.\"},\"tunes\":{}},{\"id\":\"h-matrix\",\"type\":\"header\",\"data\":{\"text\":\"A failure matrix for common RAG symptoms\",\"level\":2},\"tunes\":{}},{\"id\":\"symptom-matrix\",\"type\":\"table\",\"data\":{\"withHeadings\":true,\"stretched\":false,\"content\":[[\"Symptom\",\"Most likely layers to test first\",\"Discriminating test\"],[\"No relevant source appears\",\"Source coverage → Query → Candidate retrieval\",\"Search the corpus manually, then inspect rewritten query and unfiltered candidates\"],[\"Relevant source appears but answer is wrong\",\"Context assembly → Generation\",\"Oracle-context test with the same source reduced to decisive passages\"],[\"Answer is correct sometimes, wrong other times\",\"Ranking → Context assembly → Generation variability\",\"Repeat trials while logging retrieved set, rank, prompt context, and model output\"],[\"Answer cites the right document but overstates it\",\"Generation → Evidence attribution → Validity\",\"Grade each claim against the exact cited passage\"],[\"Old information keeps winning\",\"Ranking → Validity\u002Ffreshness\",\"Compare with recency\u002Fsupersession rules and inspect metadata\"],[\"Answer misses an exception\",\"Chunking → Context assembly\",\"Check whether rule and exception were split or truncated\"],[\"Adding more top-k makes quality worse\",\"Ranking → Context overload\",\"Ablate low-value chunks and compare with a minimal evidence set\"],[\"Changing the model fixes the answer\",\"Generation, but not necessarily retrieval\",\"Repeat with identical retrieved context across models\"],[\"Changing embeddings fixes the answer\",\"Retrieval\u002Franking\",\"Keep generator and context template constant while comparing candidate recall\"]]},\"tunes\":{}},{\"id\":\"h-metrics\",\"type\":\"header\",\"data\":{\"text\":\"Measure each layer with the metric it can actually influence\",\"level\":2},\"tunes\":{}},{\"id\":\"metrics-table\",\"type\":\"table\",\"data\":{\"withHeadings\":true,\"stretched\":false,\"content\":[[\"Layer\",\"Useful measurements\",\"What not to infer\"],[\"Source coverage\",\"Answerable-question rate, corpus coverage, ingestion completeness\",\"Do not blame embeddings for missing source material\"],[\"Candidate retrieval\",\"Recall@k, hit rate, context coverage\",\"High recall does not prove ranking quality\"],[\"Ranking\",\"MRR, NDCG, gold rank, precision@k\",\"Good ranking does not prove the generator used the evidence\"],[\"Context assembly\",\"Evidence retention, duplication, contradiction rate, token utilization\",\"Large context does not mean useful context\"],[\"Generation\",\"Correctness, completeness, task success, faithfulness\",\"Correctness alone does not prove grounding\"],[\"Evidence attribution\",\"Citation precision, citation coverage, claim support\",\"A citation count is not evidence quality\"],[\"Validity\",\"Freshness, supersession accuracy, version\u002Fjurisdiction match\",\"Relevant evidence is not automatically applicable evidence\"]]},\"tunes\":{}},{\"id\":\"h-correct-answer\",\"type\":\"header\",\"data\":{\"text\":\"A correct answer can still hide a RAG defect\",\"level\":2},\"tunes\":{}},{\"id\":\"p-correct-1\",\"type\":\"paragraph\",\"data\":{\"text\":\"The reverse problem also matters. A RAG system can produce the correct answer while retrieval is broken. The model may already know the answer from training, infer it from weak evidence, or guess correctly. If evaluation looks only at the final answer, the system can appear healthy until the question reaches information that exists only in the private corpus.\"},\"tunes\":{}},{\"id\":\"p-correct-2\",\"type\":\"paragraph\",\"data\":{\"text\":\"This is the same reliability problem that appears in agent systems more broadly: outcome correctness is not enough to prove that the execution path was reliable. For RAG, traces should preserve at least the retrieval query, candidate set, ranking, final context, answer, citations, model version, corpus\u002Findex version, and relevant filters.\"},\"tunes\":{}},{\"id\":\"internal-reliability\",\"type\":\"referralArticle\",\"data\":{\"url\":\"https:\u002F\u002Fstajic.de\u002Fblog\u002Fai-agent-reliability-why-the-final-answer-is-not-enough\",\"title\":\"AI Agent Reliability: Why the Final Answer Is Not Enough\",\"excerpt\":\"Correct output does not prove correct reasoning, safe execution, or a trustworthy system. This article extends that principle from RAG diagnosis to agent trajectories and operational assurance.\",\"ctaLabel\":\"Read the related article\"},\"tunes\":{}},{\"id\":\"h-hypotheses\",\"type\":\"header\",\"data\":{\"text\":\"Use competing hypotheses, not a favourite explanation\",\"level\":2},\"tunes\":{}},{\"id\":\"p-hyp-1\",\"type\":\"paragraph\",\"data\":{\"text\":\"If a bad answer immediately becomes “an embedding problem,” the investigation is already biased. A stronger debugging method writes down competing hypotheses before changing the system: missing source, bad query rewrite, low retrieval recall, bad reranking, context truncation, conflicting versions, generation failure, citation failure, or stale evidence.\"},\"tunes\":{}},{\"id\":\"p-hyp-2\",\"type\":\"paragraph\",\"data\":{\"text\":\"Then choose a test that would separate those hypotheses. This is more efficient than collecting more examples that support the first explanation. The same principle applies to AI-assisted technical reasoning in general: a useful diagnosis is one that survives discriminating tests, not one that merely sounds plausible.\"},\"tunes\":{}},{\"id\":\"internal-reasoning\",\"type\":\"referralArticle\",\"data\":{\"url\":\"https:\u002F\u002Fstajic.de\u002Fblog\u002Ffrom-research-protocol-to-a-general-ai-reasoning-framework\",\"title\":\"From Research Protocol to a General AI Reasoning Framework\",\"excerpt\":\"A domain-independent reasoning method for separating evidence from assumptions, testing competing hypotheses and using domain-specific validators.\",\"ctaLabel\":\"Read the reasoning framework\"},\"tunes\":{}},{\"id\":\"h-change\",\"type\":\"header\",\"data\":{\"text\":\"What would change this answer?\",\"level\":2},\"tunes\":{}},{\"id\":\"p-change-1\",\"type\":\"paragraph\",\"data\":{\"text\":\"The exact diagnostic layers change with architecture. A simple single-document RAG application may have no query rewriting, reranker, or citation layer. An agentic retrieval system may add planning, multiple searches, tool selection, memory, permissions, and iterative evidence gathering. A structured database lookup may not use chunks or embeddings at all.\"},\"tunes\":{}},{\"id\":\"p-change-2\",\"type\":\"paragraph\",\"data\":{\"text\":\"The core method still holds: identify the components that can independently change the result, construct controlled tests that replace uncertain components with known-good inputs, and measure each component using evidence appropriate to that layer.\"},\"tunes\":{}},{\"id\":\"h-limitations\",\"type\":\"header\",\"data\":{\"text\":\"Limitations\",\"level\":2},\"tunes\":{}},{\"id\":\"p-limit-1\",\"type\":\"paragraph\",\"data\":{\"text\":\"Real failures are often coupled. A weak query can reduce recall, which changes reranking, which changes context, which increases generation variance. The oracle-context test is a diagnostic shortcut, not proof that one component is solely responsible. Evaluation datasets can also be unrepresentative, and model-based graders can introduce their own errors.\"},\"tunes\":{}},{\"id\":\"p-limit-2\",\"type\":\"paragraph\",\"data\":{\"text\":\"The proposed stack is therefore best used as an investigation structure: log the pipeline, isolate variables, reproduce failures, test competing explanations, and keep end-to-end evaluation after layer-level fixes.\"},\"tunes\":{}},{\"id\":\"h-conclusion\",\"type\":\"header\",\"data\":{\"text\":\"Conclusion\",\"level\":2},\"tunes\":{}},{\"id\":\"p-conclusion-1\",\"type\":\"paragraph\",\"data\":{\"text\":\"“RAG failed” should be the beginning of the investigation, not the conclusion. A useful diagnosis identifies whether the system lacked the evidence, searched incorrectly, failed to retrieve it, ranked it badly, assembled unusable context, generated incorrectly, attributed claims poorly, or applied evidence outside its validity boundary.\"},\"tunes\":{}},{\"id\":\"p-conclusion-2\",\"type\":\"paragraph\",\"data\":{\"text\":\"The practical rule is simple: replace uncertainty with controlled evidence one layer at a time. Start with the oracle-context test. Separate retrieval-only evaluation from generation evaluation. Preserve the full trace. Then fix the component that actually failed instead of tuning the entire RAG stack by intuition.\"},\"tunes\":{}},{\"id\":\"h-faq\",\"type\":\"header\",\"data\":{\"text\":\"FAQ\",\"level\":2},\"tunes\":{}},{\"id\":\"faq\",\"type\":\"faq\",\"data\":{\"title\":\"RAG failure diagnosis\",\"items\":[{\"id\":\"faq1\",\"question\":\"How can I tell whether RAG retrieval or the LLM failed?\",\"answer\":\"Give the model a small set of known-correct evidence manually. If the answer becomes correct, investigate source coverage, query construction, retrieval, ranking, and context assembly. If the model still fails with sufficient evidence, retrieval is not the primary problem.\"},{\"id\":\"faq2\",\"question\":\"Can RAG fail even when the correct document was retrieved?\",\"answer\":\"Yes. The relevant passage can be ranked too low, truncated, separated from an exception, mixed with conflicting evidence, overwhelmed by irrelevant context, or used incorrectly by the generator.\"},{\"id\":\"faq3\",\"question\":\"Is answer correctness enough to evaluate a RAG system?\",\"answer\":\"No. A model can produce a correct answer despite weak retrieval by relying on prior model knowledge or chance. Evaluate retrieval and evidence support separately from final-answer correctness.\"},{\"id\":\"faq4\",\"question\":\"What should I log when debugging RAG?\",\"answer\":\"At minimum log the user request, transformed retrieval query, filters, candidate documents and ranks, final selected context, model and prompt version, answer, citations, corpus\u002Findex version, and timing or version metadata relevant to freshness.\"},{\"id\":\"faq5\",\"question\":\"Does increasing top-k usually fix RAG?\",\"answer\":\"Not reliably. A larger candidate or context set may improve recall, but it can also add noise, contradictions, duplicates, and context overload. Test whether the relevant evidence is missing before increasing top-k.\"}]},\"tunes\":{}},{\"id\":\"h-glossary\",\"type\":\"header\",\"data\":{\"text\":\"Glossary\",\"level\":2},\"tunes\":{}},{\"id\":\"glossary\",\"type\":\"glossary\",\"data\":{\"title\":\"Key diagnostic terms\",\"entries\":[{\"term\":\"Oracle-context test\",\"definition\":\"A controlled test in which the generator is given known-sufficient evidence directly to determine whether the dominant failure is upstream of generation.\",\"anchor\":\"oracle-context-test\"},{\"term\":\"Candidate retrieval\",\"definition\":\"The stage that selects an initial set of potentially relevant documents, chunks, records, or passages before final ranking or context assembly.\",\"anchor\":\"candidate-retrieval\"},{\"term\":\"Context assembly\",\"definition\":\"The process of converting retrieved evidence into the actual model input, including ordering, truncation, deduplication, formatting, and token-budget decisions.\",\"anchor\":\"context-assembly\"},{\"term\":\"Faithfulness\",\"definition\":\"The degree to which generated claims remain supported by the retrieved or supplied evidence rather than introducing unsupported content.\",\"anchor\":\"faithfulness\"},{\"term\":\"Context coverage\",\"definition\":\"A retrieval-oriented measure of whether selected evidence covers the information needed to answer the question.\",\"anchor\":\"context-coverage\"},{\"term\":\"Validity boundary\",\"definition\":\"The conditions under which a claim or answer remains applicable, such as time, version, jurisdiction, state, population, permissions, or source assumptions.\",\"anchor\":\"validity-boundary\"}]},\"tunes\":{}},{\"id\":\"h-sources\",\"type\":\"header\",\"data\":{\"text\":\"Primary sources and further reading\",\"level\":2},\"tunes\":{}},{\"id\":\"src-openai-rag\",\"type\":\"linkTool\",\"data\":{\"link\":\"https:\u002F\u002Fdevelopers.openai.com\u002Fapi\u002Fdocs\u002Fguides\u002Foptimizing-llm-accuracy\",\"meta\":{\"image\":{\"url\":\"\"},\"title\":\"OpenAI — Optimizing LLM Accuracy\",\"description\":\"OpenAI guidance separating retrieval failures from LLM failures in RAG applications.\"}},\"tunes\":{}},{\"id\":\"src-openai-evals\",\"type\":\"linkTool\",\"data\":{\"link\":\"https:\u002F\u002Fdevelopers.openai.com\u002Fapi\u002Fdocs\u002Fguides\u002Fevaluation-best-practices\",\"meta\":{\"image\":{\"url\":\"\"},\"title\":\"OpenAI — Evaluation Best Practices\",\"description\":\"Guidance on structured evaluation for variable AI systems and production-oriented test design.\"}},\"tunes\":{}},{\"id\":\"src-aws-rag\",\"type\":\"linkTool\",\"data\":{\"link\":\"https:\u002F\u002Fdocs.aws.amazon.com\u002Fbedrock\u002Flatest\u002Fuserguide\u002Fknowledge-base-evaluation-metrics.html\",\"meta\":{\"image\":{\"url\":\"\"},\"title\":\"Amazon Bedrock — RAG Evaluation Metrics\",\"description\":\"Documentation separating retrieve-only metrics from retrieve-and-generate metrics, including context relevance, coverage, faithfulness and citation measures.\"}},\"tunes\":{}},{\"id\":\"src-anthropic-evals\",\"type\":\"linkTool\",\"data\":{\"link\":\"https:\u002F\u002Fwww.anthropic.com\u002Fengineering\u002Fdemystifying-evals-for-ai-agents\",\"meta\":{\"image\":{\"url\":\"\"},\"title\":\"Anthropic — Demystifying Evals for AI Agents\",\"description\":\"Practical evaluation guidance on tasks, trials, graders, traces, regressions and production behaviour.\"}},\"tunes\":{}},{\"id\":\"src-google-rag\",\"type\":\"linkTool\",\"data\":{\"link\":\"https:\u002F\u002Fcloud.google.com\u002Fuse-cases\u002Fretrieval-augmented-generation\",\"meta\":{\"image\":{\"url\":\"\"},\"title\":\"Google Cloud — Retrieval-Augmented Generation\",\"description\":\"Overview of RAG architecture and the importance of relevant retrieval and grounded generation.\"}},\"tunes\":{}}],\"version\":\"2.31.6\"}",{"time":212,"blocks":213,"version":833},1790369097340,[214,222,228,236,243,248,253,258,263,268,310,315,320,325,332,337,342,347,352,357,362,367,372,377,382,387,392,397,402,407,412,417,422,427,432,437,442,447,477,484,489,521,526,531,536,541,586,591,627,632,637,642,651,656,661,666,674,679,684,689,694,699,704,709,714,719,724,750,755,782,787,797,806,815,824],{"id":215,"data":216,"type":220,"tunes":221},"8zyFXn5HD5",{"title":217,"maxLevel":218,"minLevel":219},"Contents",3,2,"tableOfContents",{},{"id":223,"data":224,"type":226,"tunes":227},"intro",{"text":225},"A RAG system returns a weak, wrong, incomplete, or unsupported answer. The usual diagnosis is “retrieval failed” or “the model hallucinated.” Both labels are too broad to be useful. A production RAG pipeline can fail before retrieval, during retrieval, while ranking, while assembling context, during generation, or after generation when evidence and validity are checked.","paragraph",{},{"id":229,"data":230,"type":234,"tunes":235},"direct",{"body":231,"title":232,"variant":233},"\u003Cstrong>Do not debug RAG as one component.\u003C\u002Fstrong> Diagnose it as a chain of independently testable layers. First determine whether the required evidence exists in an authoritative source. Then test query construction, candidate retrieval, ranking, context assembly, generation, evidence attribution, and freshness. The fastest isolation technique is an \u003Cstrong>oracle-context test\u003C\u002Fstrong>: give the generator the correct evidence manually. If the answer becomes correct, the dominant failure is upstream of generation. If it remains wrong, retrieval is not the primary problem.","Direct answer","info","callout",{},{"id":237,"data":238,"type":234,"tunes":242},"model-note",{"body":239,"title":240,"variant":241},"The RAG Failure Stack in this article is a practical diagnostic model, not a formal industry standard. Existing platforms already separate retrieval-only metrics from retrieve-and-generate metrics; this model extends that separation into a step-by-step production debugging method.","About the diagnostic model","note",{},{"id":244,"data":245,"type":42,"tunes":247},"h-not-diagnosis",{"text":246,"level":219},"Why “RAG failed” is not a diagnosis",{},{"id":249,"data":250,"type":226,"tunes":252},"p-not-1",{"text":251},"Retrieval-augmented generation combines several mechanisms: a user request is interpreted, one or more searches are constructed, candidate material is retrieved, results are filtered or reranked, selected evidence is inserted into a model context, and a model generates an answer. Production systems may add permissions, metadata filters, freshness rules, citations, query rewriting, hybrid search, tool calls, memory, and external state.",{},{"id":254,"data":255,"type":226,"tunes":257},"p-not-2",{"text":256},"A wrong final answer therefore does not tell you which component failed. The model may have received the wrong evidence. It may have received the right evidence mixed with too much noise. The evidence may be correct but stale. The source may never have contained the answer. Or the model may have ignored perfectly adequate context.",{},{"id":259,"data":260,"type":226,"tunes":262},"p-not-3",{"text":261},"OpenAI's RAG guidance already makes a fundamental distinction between retrieval failure and model failure: a system can supply the wrong context, or it can supply the right context and still generate the wrong answer. AWS similarly separates retrieve-only evaluation from retrieve-and-generate evaluation. For production diagnosis, that distinction should be taken further.",{},{"id":264,"data":265,"type":42,"tunes":267},"h-stack",{"text":266,"level":219},"The RAG Failure Stack",{},{"id":269,"data":270,"type":308,"tunes":309},"table-stack",{"content":271,"stretched":43,"withHeadings":14},[272,276,280,284,288,292,296,300,304],[273,274,275],"Layer","Question","Typical failure",[277,278,279],"1. Source coverage","Does the required evidence exist in an allowed authoritative source?","The corpus cannot answer the question at all",[281,282,283],"2. Query construction","Did the system search for the right thing?","Intent, entities, filters, language, or time constraints are lost",[285,286,287],"3. Candidate retrieval","Did the relevant evidence enter the candidate set?","Low recall; the right chunk is never retrieved",[289,290,291],"4. Ranking & filtering","Did the right evidence survive and rank high enough?","Relevant evidence is buried, filtered out, or outranked by superficially similar text",[293,294,295],"5. Context assembly","Did the model receive usable evidence?","Truncation, bad chunk boundaries, duplicates, conflicting passages, or context overload",[297,298,299],"6. Generation","Did the model use the supplied evidence correctly?","Unsupported inference, instruction failure, reasoning error, or refusal mismatch",[301,302,303],"7. Evidence attribution","Can the answer be traced to the evidence it claims to use?","Missing, weak, or incorrect citations; claims exceed retrieved support",[305,306,307],"8. Validity & freshness","Is the evidence still valid for this question now?","Correct historical evidence is reused outside its valid time, version, jurisdiction, or state","table",{},{"id":311,"data":312,"type":42,"tunes":314},"h-source",{"text":313,"level":218},"Layer 1 — Source coverage: can the system answer this at all?",{},{"id":316,"data":317,"type":226,"tunes":319},"p-source-1",{"text":318},"Before tuning embeddings, rerankers, or prompts, verify that the answer exists in the knowledge space the system is allowed to use. This sounds obvious, but many RAG failures are actually corpus failures. The requested fact may be absent, hidden in an unindexed attachment, available only in a newer document, stored in a system outside the RAG corpus, or blocked by permissions.",{},{"id":321,"data":322,"type":226,"tunes":324},"p-source-2",{"text":323},"A retrieval metric cannot recover information that was never indexed. A larger top-k cannot retrieve a document the pipeline does not contain. If the source coverage test fails, the correct fix is ingestion, source selection, permissions, or an explicit “not answerable from available evidence” behaviour.",{},{"id":326,"data":327,"type":234,"tunes":331},"source-warning",{"body":328,"title":329,"variant":330},"Teams often tune retrieval against questions that the corpus cannot actually answer. This can make the retriever better at finding related text while leaving the underlying information gap untouched.","Failure pattern","warning",{},{"id":333,"data":334,"type":42,"tunes":336},"h-query",{"text":335,"level":218},"Layer 2 — Query construction: did the system ask the corpus the right question?",{},{"id":338,"data":339,"type":226,"tunes":341},"p-query-1",{"text":340},"The user query is not always the retrieval query. Production systems rewrite questions, resolve pronouns, extract entities, translate languages, add metadata constraints, split complex questions, or generate multiple searches. Every transformation can improve retrieval, but every transformation can also destroy information.",{},{"id":343,"data":344,"type":226,"tunes":346},"p-query-2",{"text":345},"A request such as “Does the policy still apply to contractors in Germany after the September update?” contains at least an entity, a population, a jurisdiction, and a time boundary. A rewritten query that becomes “contractor policy” may retrieve semantically related text while losing the variables that decide whether the answer is valid.",{},{"id":348,"data":349,"type":42,"tunes":351},"h-retrieval",{"text":350,"level":218},"Layer 3 — Candidate retrieval: did the relevant evidence enter the set?",{},{"id":353,"data":354,"type":226,"tunes":356},"p-ret-1",{"text":355},"Candidate retrieval is primarily a recall problem. The diagnostic question is not yet whether the best result ranked first; it is whether relevant evidence appeared anywhere in the candidate pool. If the known correct source does not appear, investigate indexing, chunking, embeddings, lexical matching, metadata, hybrid search, language handling, synonyms, and query expansion.",{},{"id":358,"data":359,"type":226,"tunes":361},"p-ret-2",{"text":360},"This is where retrieval-only evaluation is valuable. AWS exposes context relevance and context coverage for retrieve-only RAG evaluation. The important production habit is to evaluate retrieval before generation so that a polished final answer cannot hide a weak candidate set.",{},{"id":363,"data":364,"type":42,"tunes":366},"h-ranking",{"text":365,"level":218},"Layer 4 — Ranking and filtering: was the right evidence discarded or buried?",{},{"id":368,"data":369,"type":226,"tunes":371},"p-rank-1",{"text":370},"A system can have good recall and still fail because the relevant evidence ranks below noisy but semantically similar material. Rerankers, recency boosts, authority weights, language preferences, tenant filters, access controls, product status filters, and deduplication all change what survives into the final context.",{},{"id":373,"data":374,"type":226,"tunes":376},"p-rank-2",{"text":375},"Debugging should therefore preserve the full candidate list, not only the final top-k. If the gold evidence was retrieved at rank 18 and a reranker removed it, the fix is not the same as a retrieval miss.",{},{"id":378,"data":379,"type":42,"tunes":381},"h-context",{"text":380,"level":218},"Layer 5 — Context assembly: did useful evidence become usable context?",{},{"id":383,"data":384,"type":226,"tunes":386},"p-ctx-1",{"text":385},"Retrieval success does not guarantee context success. Relevant chunks can be truncated, separated from their qualifiers, duplicated until they dominate the prompt, mixed with contradictory versions, or surrounded by enough irrelevant text that the decisive passage loses salience.",{},{"id":388,"data":389,"type":226,"tunes":391},"p-ctx-2",{"text":390},"Chunk boundaries are especially important. A sentence may contain the rule while the following sentence contains the exception. If they are indexed separately and only the first is retrieved, the retriever can appear relevant while the assembled context becomes misleading.",{},{"id":393,"data":394,"type":42,"tunes":396},"h-generation",{"text":395,"level":218},"Layer 6 — Generation: can the model use correct evidence correctly?",{},{"id":398,"data":399,"type":226,"tunes":401},"p-gen-1",{"text":400},"Once the system has demonstrably supplied sufficient evidence, generation becomes independently testable. The model may overgeneralize, combine incompatible passages, ignore a negative statement, fail to follow the requested answer format, invent a bridge between facts, or answer from parametric memory instead of the retrieved evidence.",{},{"id":403,"data":404,"type":226,"tunes":406},"p-gen-2",{"text":405},"This is why end-to-end correctness alone is insufficient for diagnosis. OpenAI recommends evaluation as a structured way to understand application behaviour, while Anthropic's agent-evaluation guidance emphasizes multiple trials, graders, traces, and realistic failure cases. For RAG, the generator should be tested both with normal retrieval and with controlled gold context.",{},{"id":408,"data":409,"type":42,"tunes":411},"h-evidence",{"text":410,"level":218},"Layer 7 — Evidence attribution: is the answer actually supported?",{},{"id":413,"data":414,"type":226,"tunes":416},"p-evidence-1",{"text":415},"A plausible answer with citations can still be weakly grounded. The cited document may be relevant to the topic but not support the specific claim. One sentence may be supported while another is inferred. A citation may point to a source that contradicts the answer once its conditions are read.",{},{"id":418,"data":419,"type":226,"tunes":421},"p-evidence-2",{"text":420},"Citation evaluation therefore belongs after generation. AWS distinguishes citation precision from citation coverage: whether cited passages are correctly cited and whether the answer is sufficiently supported by citations. In production, claim-level support is more useful than treating the presence of any citation as evidence quality.",{},{"id":423,"data":424,"type":42,"tunes":426},"h-validity",{"text":425,"level":218},"Layer 8 — Validity and freshness: was the evidence correct for this version of reality?",{},{"id":428,"data":429,"type":226,"tunes":431},"p-valid-1",{"text":430},"RAG can retrieve a perfectly authentic, highly relevant, faithfully quoted source and still produce a wrong answer if the source is no longer valid for the current question. Policies change. APIs are deprecated. prices move. software behaviour changes between versions. product inventory changes. permissions change. game patches change mechanics.",{},{"id":433,"data":434,"type":226,"tunes":436},"p-valid-2",{"text":435},"This is a separate failure class from hallucination. The evidence is real; its applicability is wrong. A robust system therefore needs timestamps, version or jurisdiction metadata where relevant, source authority, supersession rules, and an explicit mechanism for deciding when older evidence must be restricted or abandoned.",{},{"id":438,"data":439,"type":42,"tunes":441},"h-oracle",{"text":440,"level":219},"The fastest isolation method: the oracle-context test",{},{"id":443,"data":444,"type":226,"tunes":446},"p-oracle-1",{"text":445},"The most useful first split is simple: manually provide the generator with a small set of evidence that you know is sufficient to answer the question. Keep the task and expected answer unchanged.",{},{"id":448,"data":449,"type":475,"tunes":476},"oracle-comparison",{"rows":450,"title":464,"layout":308,"columns":465},[451,456,460],{"id":452,"label":453,"values":454},"oracle-pass","Answer becomes correct",[455,455,455],"",{"id":457,"label":458,"values":459},"oracle-fail","Answer remains wrong",[455,455,455],{"id":461,"label":462,"values":463},"oracle-partial","Answer improves but remains incomplete",[455,455,455],"Oracle-context test",[466,469,472],{"id":467,"label":468},"result","Result",{"id":470,"label":471},"meaning","Likely interpretation",{"id":473,"label":474},"next","Next diagnostic step","comparison",{},{"id":478,"data":479,"type":234,"tunes":483},"oracle-tip",{"body":480,"title":481,"variant":482},"The oracle-context test removes most of the retrieval pipeline from the experiment. It does not prove that generation is perfect, but it gives you a fast counterfactual: \u003Cstrong>what would the model do if retrieval had already succeeded?\u003C\u002Fstrong>","Why this test is powerful","tip",{},{"id":485,"data":486,"type":42,"tunes":488},"h-sequence",{"text":487,"level":219},"A production diagnostic sequence",{},{"id":490,"data":491,"type":519,"tunes":520},"diag-flow",{"steps":492,"title":517,"orientation":518},[493,496,499,502,505,508,511,514],{"label":494,"description":495},"1. Define the expected claim","Write the expected answer, allowed uncertainty, and the evidence that would justify it.",{"label":497,"description":498},"2. Verify source coverage","Confirm that authoritative and permitted evidence exists in the indexed or reachable source set.",{"label":500,"description":501},"3. Run the oracle-context test","Supply sufficient gold evidence directly to the generator and observe whether the answer becomes correct.",{"label":503,"description":504},"4. Inspect the retrieval query","Check rewrites, entities, filters, language, time constraints, decomposition, and hidden assumptions.",{"label":506,"description":507},"5. Inspect candidates before reranking","Determine whether relevant evidence was retrieved at all and record its rank.",{"label":509,"description":510},"6. Inspect ranking and context assembly","Check reranking, metadata filters, truncation, chunk boundaries, duplicates, conflicts, and top-k composition.",{"label":512,"description":513},"7. Grade generation and citations separately","Measure answer correctness, completeness, faithfulness, and claim-level evidence support.",{"label":515,"description":516},"8. Test validity boundaries","Check whether version, date, state, jurisdiction, permissions, or superseding evidence changes the answer.","Diagnose the failure from evidence to answer","auto","processFlow",{},{"id":522,"data":523,"type":42,"tunes":525},"h-one-change",{"text":524,"level":219},"Do not change three layers at once",{},{"id":527,"data":528,"type":226,"tunes":530},"p-one-1",{"text":529},"A common debugging mistake is to change embeddings, chunk sizes, top-k, prompts, and the model in one iteration. If the score improves, you do not know why. If it gets worse, you do not know which change caused the regression.",{},{"id":532,"data":533,"type":226,"tunes":535},"p-one-2",{"text":534},"Treat RAG debugging like experimental diagnosis: hold as much of the pipeline constant as possible and replace one uncertain component with a controlled input. Gold documents isolate retrieval. Gold chunks isolate chunk selection. Fixed context isolates generation. A fixed model isolates retrieval changes. A fixed corpus isolates ingestion and indexing changes.",{},{"id":537,"data":538,"type":42,"tunes":540},"h-matrix",{"text":539,"level":219},"A failure matrix for common RAG symptoms",{},{"id":542,"data":543,"type":308,"tunes":585},"symptom-matrix",{"content":544,"stretched":43,"withHeadings":14},[545,549,553,557,561,565,569,573,577,581],[546,547,548],"Symptom","Most likely layers to test first","Discriminating test",[550,551,552],"No relevant source appears","Source coverage → Query → Candidate retrieval","Search the corpus manually, then inspect rewritten query and unfiltered candidates",[554,555,556],"Relevant source appears but answer is wrong","Context assembly → Generation","Oracle-context test with the same source reduced to decisive passages",[558,559,560],"Answer is correct sometimes, wrong other times","Ranking → Context assembly → Generation variability","Repeat trials while logging retrieved set, rank, prompt context, and model output",[562,563,564],"Answer cites the right document but overstates it","Generation → Evidence attribution → Validity","Grade each claim against the exact cited passage",[566,567,568],"Old information keeps winning","Ranking → Validity\u002Ffreshness","Compare with recency\u002Fsupersession rules and inspect metadata",[570,571,572],"Answer misses an exception","Chunking → Context assembly","Check whether rule and exception were split or truncated",[574,575,576],"Adding more top-k makes quality worse","Ranking → Context overload","Ablate low-value chunks and compare with a minimal evidence set",[578,579,580],"Changing the model fixes the answer","Generation, but not necessarily retrieval","Repeat with identical retrieved context across models",[582,583,584],"Changing embeddings fixes the answer","Retrieval\u002Franking","Keep generator and context template constant while comparing candidate recall",{},{"id":587,"data":588,"type":42,"tunes":590},"h-metrics",{"text":589,"level":219},"Measure each layer with the metric it can actually influence",{},{"id":592,"data":593,"type":308,"tunes":626},"metrics-table",{"content":594,"stretched":43,"withHeadings":14},[595,598,602,606,610,614,618,622],[273,596,597],"Useful measurements","What not to infer",[599,600,601],"Source coverage","Answerable-question rate, corpus coverage, ingestion completeness","Do not blame embeddings for missing source material",[603,604,605],"Candidate retrieval","Recall@k, hit rate, context coverage","High recall does not prove ranking quality",[607,608,609],"Ranking","MRR, NDCG, gold rank, precision@k","Good ranking does not prove the generator used the evidence",[611,612,613],"Context assembly","Evidence retention, duplication, contradiction rate, token utilization","Large context does not mean useful context",[615,616,617],"Generation","Correctness, completeness, task success, faithfulness","Correctness alone does not prove grounding",[619,620,621],"Evidence attribution","Citation precision, citation coverage, claim support","A citation count is not evidence quality",[623,624,625],"Validity","Freshness, supersession accuracy, version\u002Fjurisdiction match","Relevant evidence is not automatically applicable evidence",{},{"id":628,"data":629,"type":42,"tunes":631},"h-correct-answer",{"text":630,"level":219},"A correct answer can still hide a RAG defect",{},{"id":633,"data":634,"type":226,"tunes":636},"p-correct-1",{"text":635},"The reverse problem also matters. A RAG system can produce the correct answer while retrieval is broken. The model may already know the answer from training, infer it from weak evidence, or guess correctly. If evaluation looks only at the final answer, the system can appear healthy until the question reaches information that exists only in the private corpus.",{},{"id":638,"data":639,"type":226,"tunes":641},"p-correct-2",{"text":640},"This is the same reliability problem that appears in agent systems more broadly: outcome correctness is not enough to prove that the execution path was reliable. For RAG, traces should preserve at least the retrieval query, candidate set, ranking, final context, answer, citations, model version, corpus\u002Findex version, and relevant filters.",{},{"id":643,"data":644,"type":649,"tunes":650},"internal-reliability",{"url":645,"title":646,"excerpt":647,"ctaLabel":648},"https:\u002F\u002Fstajic.de\u002Fblog\u002Fai-agent-reliability-why-the-final-answer-is-not-enough","AI Agent Reliability: Why the Final Answer Is Not Enough","Correct output does not prove correct reasoning, safe execution, or a trustworthy system. This article extends that principle from RAG diagnosis to agent trajectories and operational assurance.","Read the related article","referralArticle",{},{"id":652,"data":653,"type":42,"tunes":655},"h-hypotheses",{"text":654,"level":219},"Use competing hypotheses, not a favourite explanation",{},{"id":657,"data":658,"type":226,"tunes":660},"p-hyp-1",{"text":659},"If a bad answer immediately becomes “an embedding problem,” the investigation is already biased. A stronger debugging method writes down competing hypotheses before changing the system: missing source, bad query rewrite, low retrieval recall, bad reranking, context truncation, conflicting versions, generation failure, citation failure, or stale evidence.",{},{"id":662,"data":663,"type":226,"tunes":665},"p-hyp-2",{"text":664},"Then choose a test that would separate those hypotheses. This is more efficient than collecting more examples that support the first explanation. The same principle applies to AI-assisted technical reasoning in general: a useful diagnosis is one that survives discriminating tests, not one that merely sounds plausible.",{},{"id":667,"data":668,"type":649,"tunes":673},"internal-reasoning",{"url":669,"title":670,"excerpt":671,"ctaLabel":672},"https:\u002F\u002Fstajic.de\u002Fblog\u002Ffrom-research-protocol-to-a-general-ai-reasoning-framework","From Research Protocol to a General AI Reasoning Framework","A domain-independent reasoning method for separating evidence from assumptions, testing competing hypotheses and using domain-specific validators.","Read the reasoning framework",{},{"id":675,"data":676,"type":42,"tunes":678},"h-change",{"text":677,"level":219},"What would change this answer?",{},{"id":680,"data":681,"type":226,"tunes":683},"p-change-1",{"text":682},"The exact diagnostic layers change with architecture. A simple single-document RAG application may have no query rewriting, reranker, or citation layer. An agentic retrieval system may add planning, multiple searches, tool selection, memory, permissions, and iterative evidence gathering. A structured database lookup may not use chunks or embeddings at all.",{},{"id":685,"data":686,"type":226,"tunes":688},"p-change-2",{"text":687},"The core method still holds: identify the components that can independently change the result, construct controlled tests that replace uncertain components with known-good inputs, and measure each component using evidence appropriate to that layer.",{},{"id":690,"data":691,"type":42,"tunes":693},"h-limitations",{"text":692,"level":219},"Limitations",{},{"id":695,"data":696,"type":226,"tunes":698},"p-limit-1",{"text":697},"Real failures are often coupled. A weak query can reduce recall, which changes reranking, which changes context, which increases generation variance. The oracle-context test is a diagnostic shortcut, not proof that one component is solely responsible. Evaluation datasets can also be unrepresentative, and model-based graders can introduce their own errors.",{},{"id":700,"data":701,"type":226,"tunes":703},"p-limit-2",{"text":702},"The proposed stack is therefore best used as an investigation structure: log the pipeline, isolate variables, reproduce failures, test competing explanations, and keep end-to-end evaluation after layer-level fixes.",{},{"id":705,"data":706,"type":42,"tunes":708},"h-conclusion",{"text":707,"level":219},"Conclusion",{},{"id":710,"data":711,"type":226,"tunes":713},"p-conclusion-1",{"text":712},"“RAG failed” should be the beginning of the investigation, not the conclusion. A useful diagnosis identifies whether the system lacked the evidence, searched incorrectly, failed to retrieve it, ranked it badly, assembled unusable context, generated incorrectly, attributed claims poorly, or applied evidence outside its validity boundary.",{},{"id":715,"data":716,"type":226,"tunes":718},"p-conclusion-2",{"text":717},"The practical rule is simple: replace uncertainty with controlled evidence one layer at a time. Start with the oracle-context test. Separate retrieval-only evaluation from generation evaluation. Preserve the full trace. Then fix the component that actually failed instead of tuning the entire RAG stack by intuition.",{},{"id":720,"data":721,"type":42,"tunes":723},"h-faq",{"text":722,"level":219},"FAQ",{},{"id":725,"data":726,"type":725,"tunes":749},"faq",{"items":727,"title":748},[728,732,736,740,744],{"id":729,"answer":730,"question":731},"faq1","Give the model a small set of known-correct evidence manually. If the answer becomes correct, investigate source coverage, query construction, retrieval, ranking, and context assembly. If the model still fails with sufficient evidence, retrieval is not the primary problem.","How can I tell whether RAG retrieval or the LLM failed?",{"id":733,"answer":734,"question":735},"faq2","Yes. The relevant passage can be ranked too low, truncated, separated from an exception, mixed with conflicting evidence, overwhelmed by irrelevant context, or used incorrectly by the generator.","Can RAG fail even when the correct document was retrieved?",{"id":737,"answer":738,"question":739},"faq3","No. A model can produce a correct answer despite weak retrieval by relying on prior model knowledge or chance. Evaluate retrieval and evidence support separately from final-answer correctness.","Is answer correctness enough to evaluate a RAG system?",{"id":741,"answer":742,"question":743},"faq4","At minimum log the user request, transformed retrieval query, filters, candidate documents and ranks, final selected context, model and prompt version, answer, citations, corpus\u002Findex version, and timing or version metadata relevant to freshness.","What should I log when debugging RAG?",{"id":745,"answer":746,"question":747},"faq5","Not reliably. A larger candidate or context set may improve recall, but it can also add noise, contradictions, duplicates, and context overload. Test whether the relevant evidence is missing before increasing top-k.","Does increasing top-k usually fix RAG?","RAG failure diagnosis",{},{"id":751,"data":752,"type":42,"tunes":754},"h-glossary",{"text":753,"level":219},"Glossary",{},{"id":756,"data":757,"type":756,"tunes":781},"glossary",{"title":758,"entries":759},"Key diagnostic terms",[760,763,766,769,773,777],{"term":464,"anchor":761,"definition":762},"oracle-context-test","A controlled test in which the generator is given known-sufficient evidence directly to determine whether the dominant failure is upstream of generation.",{"term":603,"anchor":764,"definition":765},"candidate-retrieval","The stage that selects an initial set of potentially relevant documents, chunks, records, or passages before final ranking or context assembly.",{"term":611,"anchor":767,"definition":768},"context-assembly","The process of converting retrieved evidence into the actual model input, including ordering, truncation, deduplication, formatting, and token-budget decisions.",{"term":770,"anchor":771,"definition":772},"Faithfulness","faithfulness","The degree to which generated claims remain supported by the retrieved or supplied evidence rather than introducing unsupported content.",{"term":774,"anchor":775,"definition":776},"Context coverage","context-coverage","A retrieval-oriented measure of whether selected evidence covers the information needed to answer the question.",{"term":778,"anchor":779,"definition":780},"Validity boundary","validity-boundary","The conditions under which a claim or answer remains applicable, such as time, version, jurisdiction, state, population, permissions, or source assumptions.",{},{"id":783,"data":784,"type":42,"tunes":786},"h-sources",{"text":785,"level":219},"Primary sources and further reading",{},{"id":788,"data":789,"type":795,"tunes":796},"src-openai-rag",{"link":790,"meta":791},"https:\u002F\u002Fdevelopers.openai.com\u002Fapi\u002Fdocs\u002Fguides\u002Foptimizing-llm-accuracy",{"image":792,"title":793,"description":794},{"url":455},"OpenAI — Optimizing LLM Accuracy","OpenAI guidance separating retrieval failures from LLM failures in RAG applications.","linkTool",{},{"id":798,"data":799,"type":795,"tunes":805},"src-openai-evals",{"link":800,"meta":801},"https:\u002F\u002Fdevelopers.openai.com\u002Fapi\u002Fdocs\u002Fguides\u002Fevaluation-best-practices",{"image":802,"title":803,"description":804},{"url":455},"OpenAI — Evaluation Best Practices","Guidance on structured evaluation for variable AI systems and production-oriented test design.",{},{"id":807,"data":808,"type":795,"tunes":814},"src-aws-rag",{"link":809,"meta":810},"https:\u002F\u002Fdocs.aws.amazon.com\u002Fbedrock\u002Flatest\u002Fuserguide\u002Fknowledge-base-evaluation-metrics.html",{"image":811,"title":812,"description":813},{"url":455},"Amazon Bedrock — RAG Evaluation Metrics","Documentation separating retrieve-only metrics from retrieve-and-generate metrics, including context relevance, coverage, faithfulness and citation measures.",{},{"id":816,"data":817,"type":795,"tunes":823},"src-anthropic-evals",{"link":818,"meta":819},"https:\u002F\u002Fwww.anthropic.com\u002Fengineering\u002Fdemystifying-evals-for-ai-agents",{"image":820,"title":821,"description":822},{"url":455},"Anthropic — Demystifying Evals for AI Agents","Practical evaluation guidance on tasks, trials, graders, traces, regressions and production behaviour.",{},{"id":825,"data":826,"type":795,"tunes":832},"src-google-rag",{"link":827,"meta":828},"https:\u002F\u002Fcloud.google.com\u002Fuse-cases\u002Fretrieval-augmented-generation",{"image":829,"title":830,"description":831},{"url":455},"Google Cloud — Retrieval-Augmented Generation","Overview of RAG architecture and the importance of relevant retrieval and grounded generation.",{},"2.31.6","When a RAG answer is wrong, blaming retrieval or the model is too vague. This diagnostic method isolates source coverage, query construction, retrieval, ranking, context assembly, generation, evidence attribution, and freshness—so the actual failure can be reproduced and fixed.","\u002Fuploads\u002F2026\u002F09\u002Frag-failed-but-which-layer-actually-failed-a-diagnostic-method-1790350847177-pior4c.webp","rag-failed-but-which-layer-actually-failed-a-diagnostic-method-1790350847177-pior4c","PUBLISHED","2026-09-24T19:39:00.000Z","2026-09-25T15:39:19.132Z","2026-09-25T20:46:25.690Z",{"en":842,"de":843,"sr":844,"es":845,"fr":846,"it":847,"ru":848,"zh":849},"\u002Fblog\u002Frag-failed-but-which-layer-actually-failed-a-diagnostic-method","\u002Fde\u002Fblog\u002Frag-failed-but-which-layer-actually-failed-a-diagnostic-method","\u002Fsr\u002Fblog\u002Frag-failed-but-which-layer-actually-failed-a-diagnostic-method","\u002Fes\u002Fblog\u002Frag-failed-but-which-layer-actually-failed-a-diagnostic-method","\u002Ffr\u002Fblog\u002Frag-failed-but-which-layer-actually-failed-a-diagnostic-method","\u002Fit\u002Fblog\u002Frag-failed-but-which-layer-actually-failed-a-diagnostic-method","\u002Fru\u002Fblog\u002Frag-failed-but-which-layer-actually-failed-a-diagnostic-method","\u002Fzh\u002Fblog\u002Frag-failed-but-which-layer-actually-failed-a-diagnostic-method",[851,855,859],{"id":852,"name":853,"slug":854},58,"Evaluation & Quality Gates","evaluation",{"id":856,"name":857,"slug":858},89,"Evaluation Harness","evaluation-harness",{"id":860,"name":861,"slug":862},85,"Quality Gates","quality-gates",{"id":864,"login":865,"email":866,"displayName":867},"20","rooth8233","aleksandar@stajic.de","Aleksandar Stajić",[869],{"lang":7,"title":208,"content":210,"contentJson":870,"excerpt":834},{"time":212,"blocks":871,"version":833},[872,875,878,881,884,887,890,893,896,899,912,915,918,921,924,927,930,933,936,939,942,945,948,951,954,957,960,963,966,969,972,975,978,981,984,987,990,993,1007,1010,1013,1025,1028,1031,1034,1037,1051,1054,1066,1069,1072,1075,1078,1081,1084,1087,1090,1093,1096,1099,1102,1105,1108,1111,1114,1117,1120,1129,1132,1142,1145,1150,1155,1160,1165],{"id":215,"data":873,"type":220,"tunes":874},{"title":217,"maxLevel":218,"minLevel":219},{},{"id":223,"data":876,"type":226,"tunes":877},{"text":225},{},{"id":229,"data":879,"type":234,"tunes":880},{"body":231,"title":232,"variant":233},{},{"id":237,"data":882,"type":234,"tunes":883},{"body":239,"title":240,"variant":241},{},{"id":244,"data":885,"type":42,"tunes":886},{"text":246,"level":219},{},{"id":249,"data":888,"type":226,"tunes":889},{"text":251},{},{"id":254,"data":891,"type":226,"tunes":892},{"text":256},{},{"id":259,"data":894,"type":226,"tunes":895},{"text":261},{},{"id":264,"data":897,"type":42,"tunes":898},{"text":266,"level":219},{},{"id":269,"data":900,"type":308,"tunes":911},{"content":901,"stretched":43,"withHeadings":14},[902,903,904,905,906,907,908,909,910],[273,274,275],[277,278,279],[281,282,283],[285,286,287],[289,290,291],[293,294,295],[297,298,299],[301,302,303],[305,306,307],{},{"id":311,"data":913,"type":42,"tunes":914},{"text":313,"level":218},{},{"id":316,"data":916,"type":226,"tunes":917},{"text":318},{},{"id":321,"data":919,"type":226,"tunes":920},{"text":323},{},{"id":326,"data":922,"type":234,"tunes":923},{"body":328,"title":329,"variant":330},{},{"id":333,"data":925,"type":42,"tunes":926},{"text":335,"level":218},{},{"id":338,"data":928,"type":226,"tunes":929},{"text":340},{},{"id":343,"data":931,"type":226,"tunes":932},{"text":345},{},{"id":348,"data":934,"type":42,"tunes":935},{"text":350,"level":218},{},{"id":353,"data":937,"type":226,"tunes":938},{"text":355},{},{"id":358,"data":940,"type":226,"tunes":941},{"text":360},{},{"id":363,"data":943,"type":42,"tunes":944},{"text":365,"level":218},{},{"id":368,"data":946,"type":226,"tunes":947},{"text":370},{},{"id":373,"data":949,"type":226,"tunes":950},{"text":375},{},{"id":378,"data":952,"type":42,"tunes":953},{"text":380,"level":218},{},{"id":383,"data":955,"type":226,"tunes":956},{"text":385},{},{"id":388,"data":958,"type":226,"tunes":959},{"text":390},{},{"id":393,"data":961,"type":42,"tunes":962},{"text":395,"level":218},{},{"id":398,"data":964,"type":226,"tunes":965},{"text":400},{},{"id":403,"data":967,"type":226,"tunes":968},{"text":405},{},{"id":408,"data":970,"type":42,"tunes":971},{"text":410,"level":218},{},{"id":413,"data":973,"type":226,"tunes":974},{"text":415},{},{"id":418,"data":976,"type":226,"tunes":977},{"text":420},{},{"id":423,"data":979,"type":42,"tunes":980},{"text":425,"level":218},{},{"id":428,"data":982,"type":226,"tunes":983},{"text":430},{},{"id":433,"data":985,"type":226,"tunes":986},{"text":435},{},{"id":438,"data":988,"type":42,"tunes":989},{"text":440,"level":219},{},{"id":443,"data":991,"type":226,"tunes":992},{"text":445},{},{"id":448,"data":994,"type":475,"tunes":1006},{"rows":995,"title":464,"layout":308,"columns":1002},[996,998,1000],{"id":452,"label":453,"values":997},[455,455,455],{"id":457,"label":458,"values":999},[455,455,455],{"id":461,"label":462,"values":1001},[455,455,455],[1003,1004,1005],{"id":467,"label":468},{"id":470,"label":471},{"id":473,"label":474},{},{"id":478,"data":1008,"type":234,"tunes":1009},{"body":480,"title":481,"variant":482},{},{"id":485,"data":1011,"type":42,"tunes":1012},{"text":487,"level":219},{},{"id":490,"data":1014,"type":519,"tunes":1024},{"steps":1015,"title":517,"orientation":518},[1016,1017,1018,1019,1020,1021,1022,1023],{"label":494,"description":495},{"label":497,"description":498},{"label":500,"description":501},{"label":503,"description":504},{"label":506,"description":507},{"label":509,"description":510},{"label":512,"description":513},{"label":515,"description":516},{},{"id":522,"data":1026,"type":42,"tunes":1027},{"text":524,"level":219},{},{"id":527,"data":1029,"type":226,"tunes":1030},{"text":529},{},{"id":532,"data":1032,"type":226,"tunes":1033},{"text":534},{},{"id":537,"data":1035,"type":42,"tunes":1036},{"text":539,"level":219},{},{"id":542,"data":1038,"type":308,"tunes":1050},{"content":1039,"stretched":43,"withHeadings":14},[1040,1041,1042,1043,1044,1045,1046,1047,1048,1049],[546,547,548],[550,551,552],[554,555,556],[558,559,560],[562,563,564],[566,567,568],[570,571,572],[574,575,576],[578,579,580],[582,583,584],{},{"id":587,"data":1052,"type":42,"tunes":1053},{"text":589,"level":219},{},{"id":592,"data":1055,"type":308,"tunes":1065},{"content":1056,"stretched":43,"withHeadings":14},[1057,1058,1059,1060,1061,1062,1063,1064],[273,596,597],[599,600,601],[603,604,605],[607,608,609],[611,612,613],[615,616,617],[619,620,621],[623,624,625],{},{"id":628,"data":1067,"type":42,"tunes":1068},{"text":630,"level":219},{},{"id":633,"data":1070,"type":226,"tunes":1071},{"text":635},{},{"id":638,"data":1073,"type":226,"tunes":1074},{"text":640},{},{"id":643,"data":1076,"type":649,"tunes":1077},{"url":645,"title":646,"excerpt":647,"ctaLabel":648},{},{"id":652,"data":1079,"type":42,"tunes":1080},{"text":654,"level":219},{},{"id":657,"data":1082,"type":226,"tunes":1083},{"text":659},{},{"id":662,"data":1085,"type":226,"tunes":1086},{"text":664},{},{"id":667,"data":1088,"type":649,"tunes":1089},{"url":669,"title":670,"excerpt":671,"ctaLabel":672},{},{"id":675,"data":1091,"type":42,"tunes":1092},{"text":677,"level":219},{},{"id":680,"data":1094,"type":226,"tunes":1095},{"text":682},{},{"id":685,"data":1097,"type":226,"tunes":1098},{"text":687},{},{"id":690,"data":1100,"type":42,"tunes":1101},{"text":692,"level":219},{},{"id":695,"data":1103,"type":226,"tunes":1104},{"text":697},{},{"id":700,"data":1106,"type":226,"tunes":1107},{"text":702},{},{"id":705,"data":1109,"type":42,"tunes":1110},{"text":707,"level":219},{},{"id":710,"data":1112,"type":226,"tunes":1113},{"text":712},{},{"id":715,"data":1115,"type":226,"tunes":1116},{"text":717},{},{"id":720,"data":1118,"type":42,"tunes":1119},{"text":722,"level":219},{},{"id":725,"data":1121,"type":725,"tunes":1128},{"items":1122,"title":748},[1123,1124,1125,1126,1127],{"id":729,"answer":730,"question":731},{"id":733,"answer":734,"question":735},{"id":737,"answer":738,"question":739},{"id":741,"answer":742,"question":743},{"id":745,"answer":746,"question":747},{},{"id":751,"data":1130,"type":42,"tunes":1131},{"text":753,"level":219},{},{"id":756,"data":1133,"type":756,"tunes":1141},{"title":758,"entries":1134},[1135,1136,1137,1138,1139,1140],{"term":464,"anchor":761,"definition":762},{"term":603,"anchor":764,"definition":765},{"term":611,"anchor":767,"definition":768},{"term":770,"anchor":771,"definition":772},{"term":774,"anchor":775,"definition":776},{"term":778,"anchor":779,"definition":780},{},{"id":783,"data":1143,"type":42,"tunes":1144},{"text":785,"level":219},{},{"id":788,"data":1146,"type":795,"tunes":1149},{"link":790,"meta":1147},{"image":1148,"title":793,"description":794},{"url":455},{},{"id":798,"data":1151,"type":795,"tunes":1154},{"link":800,"meta":1152},{"image":1153,"title":803,"description":804},{"url":455},{},{"id":807,"data":1156,"type":795,"tunes":1159},{"link":809,"meta":1157},{"image":1158,"title":812,"description":813},{"url":455},{},{"id":816,"data":1161,"type":795,"tunes":1164},{"link":818,"meta":1162},{"image":1163,"title":821,"description":822},{"url":455},{},{"id":825,"data":1166,"type":795,"tunes":1169},{"link":827,"meta":1167},{"image":1168,"title":830,"description":831},{"url":455},{},"Post erfolgreich abgerufen",{"items":1172,"source":1227,"manualIds":1228,"manualMatchedIds":1229},[1173,1180,1186,1193,1200,1207,1214,1220],{"id":1174,"slug":1175,"title":1176,"excerpt":1177,"featuredImage":1178,"publishedAt":1179},"478","what-is-rag-the-simplest-explanation-of-how-it-works","What Is RAG? The Simplest Explanation of How It Works","RAG sounds complicated, but the idea is simple: before an AI answers, it first looks up useful information from a knowledge source and gives that information to the language model. This guide explains RAG, LLMs, state, memory and tools using one simple mental model.","\u002Fuploads\u002F2026\u002F09\u002Fwhat-is-rag-the-simplest-explanation-of-how-it-works-1790377492124-khjagt.webp","2026-09-25T19:03:00.000Z",{"id":1181,"slug":1182,"title":646,"excerpt":1183,"featuredImage":1184,"publishedAt":1185},"460","ai-agent-reliability-why-the-final-answer-is-not-enough","Correct output does not prove correct reasoning, safe execution, or a trustworthy system.","\u002Fuploads\u002F2026\u002F09\u002Fai-agent-reliability-why-the-final-answer-is-not-enough-1788955466306-pl0qhz.webp","2026-09-09T04:01:00.000Z",{"id":1187,"slug":1188,"title":1189,"excerpt":1190,"featuredImage":1191,"publishedAt":1192},"468","ai-agent-memory-is-not-rag-how-to-separate-memory-retrieval-state-and-context","AI Agent Memory Is Not RAG: How to Separate Memory, Retrieval, State and Context","Agent memory, RAG, state, and context are often used as if they were interchangeable. They are not. This practical architecture model separates the four layers, shows where each belongs, and explains what breaks when systems collapse them into one.","\u002Fuploads\u002F2026\u002F09\u002Fai-agent-memory-is-not-rag-how-to-separate-memory-retrieval-state-and-context-1790350560308-np0xy6.webp","2026-09-25T11:34:00.000Z",{"id":1194,"slug":1195,"title":1196,"excerpt":1197,"featuredImage":1198,"publishedAt":1199},"477","computer-use-agents-why-a-successful-demo-can-still-be-an-unreliable-system","Computer-Use Agents: Why a Successful Demo Can Still Be an Unreliable System","Computer-use agents can now complete impressive browser and desktop workflows, but one successful run proves capability—not reliability. This article shows how to test repeatability, environmental robustness, long-horizon control, state awareness, outcome verification, and safe goal handling.","\u002Fuploads\u002F2026\u002F09\u002Fcomputer-use-agents-why-a-successful-demo-can-still-be-an-unreliable-system-1790352854690-75qnrg.webp","2026-09-25T12:13:00.000Z",{"id":1201,"slug":1202,"title":1203,"excerpt":1204,"featuredImage":1205,"publishedAt":1206},"475","managed-agent-harness-vs-self-hosted-agent-loop-what-you-gain-what-you-lose","Managed Agent Harness vs Self-Hosted Agent Loop: What You Gain, What You Lose","“Self-hosted agent” can mean very different architectures. This guide separates the managed harness, self-hosted execution environment, and fully self-operated agent loop—and shows which control boundary teams actually need.","\u002Fuploads\u002F2026\u002F09\u002Fmanaged-agent-harness-vs-self-hosted-agent-loop-what-you-gain-what-you-lose-1790352403475-kj10jh.webp","2026-09-25T12:05:00.000Z",{"id":1208,"slug":1209,"title":1210,"excerpt":1211,"featuredImage":1212,"publishedAt":1213},"459","ollama-is-not-the-product-building-production-ready-open-llm-applications","Ollama Is Not the Product: Building Production-Ready Open-LLM Applications","Running a local model with Ollama is easy. Building a production-ready Open-LLM application is harder: it requires RAG, access control, provider abstraction, evaluation, logging, deployment discipline and a controlled application layer around the model.\n","\u002Fuploads\u002F2026\u002F06\u002Follama-is-not-the-product-building-production-ready-open-llm-applications-1782679361640-h0usqf.webp","2026-06-28T16:39:00.000Z",{"id":1215,"slug":858,"title":1216,"excerpt":1217,"featuredImage":1218,"publishedAt":1219},"434","Comprehensive Guide to Evaluation Harness: Mastering LLM Performance Evaluation","This guide provides a detailed walkthrough of Evaluation Harness, an essential framework for rigorously assessing large language model (LLM) capabilities in enterprise LLMOps pipelines. Learn setup, best practices, and advanced techniques to ensure reliable model benchmarking and optimization.","\u002Fuploads\u002F2026\u002F04\u002Fevaluation-harness-1775466944495-4s0xv2.webp","2026-03-01T17:50:00.000Z",{"id":1221,"slug":1222,"title":1223,"excerpt":1224,"featuredImage":1225,"publishedAt":1226},"364","tipps-fuer-die-verbesserung-der-seo-suchmaschinenoptimierung","Mastering the SEO Workflow: Essential Optimization Strategies for Organic Growth","A structured SEO workflow is crucial for sustainable organic growth. Learn the ten foundational strategies, from keyword research and technical optimization to content quality and performance analysis.","\u002Fuploads\u002F2026\u002F03\u002Ftipps-fuer-die-verbesserung-der-seo-suchmaschinenoptimierung-1774866098131-hwkzrg.webp","2024-01-26T06:35:00.000Z","fallback",[],[]]