Skip to content

Standard Reader

Tag

Embeddings

Articles and publications tagged Embeddings across the Atmosphere.

14articles5publications

Articles

Publications

Today I LearnedAug 6, 2026

Learned vector representations of meaning and similarity

artificial-intelligencemachine-learning

Sahil Kapoor's PlaybookMay 12, 2026

pgvector is the popular PostgreSQL extension that adds a vector column type, distance operators, and approximate nearest-neighbour indexes for similarity search over embeddings. It turns any PostgreSQL database into a vector database, removing the operational overhead of running a separate ANN store for many RAG and recommendation workloads. What it provides * vector(N) column type. Stores fixed-dimensional float vectors. * Distance operators. L2 distance (<->), inner product (<#>), cosine

PostgreSQLVector Database

Sahil Kapoor's PlaybookMay 12, 2026

Fine-tuning is the process of taking a pretrained model and continuing to train it on a target dataset, so its weights adapt to a specific task, domain, style, or output format. It is the standard way to customise a model when prompting alone is not enough. Variants * Full fine-tuning. Updates every parameter. Maximum capacity, maximum cost. * LoRA (Low-Rank Adaptation). Trains a small low-rank update on top of frozen weights. Fast, cheap, and the result is a small adapter file. * QLoRA. L

Sahil Kapoor's PlaybookMay 12, 2026

A reranker is a model that takes an initial set of retrieved candidates and re-orders them to improve precision. In a typical retrieval pipeline, a fast first-stage retriever returns 50 to 200 candidates, and a slower but more accurate reranker scores each (query, candidate) pair and returns the top 5 to 10. How it works Most rerankers are cross-encoders: a transformer model that takes the query and the candidate as a single concatenated input and outputs a relevance score. This differs from

Sahil Kapoor's PlaybookMay 12, 2026

Chunking is the process of splitting documents into smaller passages before embedding them for retrieval. Chunk size and boundaries directly determine what a retrieval system can find: a chunk that is too large blurs the meaning of its embedding, and a chunk that is too small lacks the context to answer most questions. Common strategies * Fixed-size character or token splitting. Cuts every N characters or tokens. Simple but ignores semantic boundaries. * Recursive character splitting. Tries

Sahil Kapoor's PlaybookMay 12, 2026

Vector Database

A vector database is a storage system optimised for similarity search over high-dimensional vectors, typically embeddings. Given a query vector, it returns the closest vectors in the corpus using an approximate nearest-neighbour index, alongside any associated metadata. How it works Vector databases index embeddings using approximate nearest-neighbour (ANN) algorithms such as HNSW, IVF, ScaNN, or DiskANN. These trade a small amount of recall for orders-of-magnitude better latency than a brute

EmbeddingsRedis

fuzzy find my breadcrumbs

atprotoembeddings

plyr.fmFeb 12, 2026

sounds like a bunch of semantics

luminary.blogSep 4, 2025

What is Matryoshka Representation Learning (MRL)?

Nesting Power and Flexibility into ML Embeddings

luminary.blogAug 4, 2025

What are Positional Embeddings?

The mathematical technique that teaches AI models where each word sits in a sequence.

luminary.blogJul 31, 2025

Words, Tokens and Embeddings

How language models convert token IDs into meaningful vector representations that capture semantic relationships.

luminary.blogMay 11, 2025

Embedding Selection for RAG Systems

At the heart of every effective RAG implementation lies a crucial decision: which embedding model to use.

EJ FoxMar 25, 2025

Mapping Novel Territories: Untangling MCP Servers

In which we the author gets his bearings by doing one of the things he knows best; making a map

embeddingscartography

luminary.blogMar 16, 2025

What are Word Embeddings?

Word embeddings are a fundamental concept in Natural Language Processing (NLP), enabling machines to understand and process human language effectively.

You've reached the end.