A bit more detail
It works by converting words, images, or documents into long strings of numbers called embeddings. Systems calculate the distance between these points to find the closest matches in meaning. It is widely used in generative AI to retrieve relevant background facts.