Grounding a model in your own data: the pipeline, the failure modes, and the frameworks.
Sign in to track your progress across sections.
17 articles
Structural, recursive and semantic splitting compared, with the sizes to start from for short lookups and for prose. Chunk size sets a ceiling on what retrieval can return, and several common mistakes lower it further.
Prepending model-generated context to each chunk before embedding, and the measured drop in failed retrievals it produced.
Pinecone walks through fixed, recursive, and document-aware splitting with concrete size and overlap recommendations.
LangChain's concept doc on the splitter family — what each one optimises for and when to reach for it.
Listed inStoring in a Vector DBRAGon
A namespace is a hard partition, not a tag. Changing the embedding model invalidates every vector, so plan the rebuild before you need it.
Listed inLlamaIndexRAGon
Documents and Nodes, and a query stage that separates retrieval, postprocessing and response synthesis. Postprocessing is the step hand-written pipelines usually skip, and it is where reranking and filtering belong.
Listed inRAG Use CasesRAGon
Four problem shapes retrieval fits and four it does not, why counting questions fail quietly, and how a million-token window moved the threshold.
Listed inRAG with Dynamic FiltersRAGon
A forgotten metadata filter returns everyone's documents; a forgotten namespace returns an error. When the filter is a tenancy boundary, that asymmetry is the whole argument for a partition.
Listed inRetrieval ProcessRAGon
Query rewriting, hybrid search, reranking, and how many chunks to actually pass.
Listed inEmbedding DocumentsRAGon
Turning chunks into vectors, with metadata that survives to query time.
Listed inRAG vs Fine-tuningRAGon
Knowledge that changes vs behaviour that doesn't — picking the right tool.
Listed inWhat is RAG?RAGon
Retrieval-augmented generation: fetch relevant context, then answer from it.
Listed inGenerationRAGon
Composing retrieved context into a prompt that forces grounded answers.