AI Flashcards
1
Q
what are embeddings used for in openAI?
A
Embeddings turn words into numerical vectors that make it easier to perform AI operation. an example is cat and dog are similar enough that they might get numbers close to each other. embeddings are used for search, sentiment analysis, and lots of other AI operations.
you can use sentences and cosign similarity to find the closest sentence, regardless of word matches like you would need with ElasticSearch.
2
Q
cosign similarity?
A
measure the similarity between 2 vectors in multidimensional space. it’s a measure between -1 and 1, and a fast way to calculate the similarity between words and sentences to quickly find context in large documents. it’s also fast.
