Embedding

Embeddings convert complex data like text and images into high-dimensional vectors to measure semantic similarity.

Embedding technology transforms discrete data points into continuous vector spaces, allowing machines to calculate the 'distance' between concepts. By mapping words, images, or audio into arrays of numbers (often 1,536 dimensions or more), models like text-embedding-3-large can identify that 'king' and 'queen' are semantically closer than 'king' and 'keyboard.' This mathematical representation is the backbone of modern AI, powering everything from semantic search and recommendation engines to retrieval-augmented generation (RAG) for large language models.

https://platform.openai.com/docs/guides/embeddings