`text-embedding-ada-002`

[curt.kennedy]

Technically you are still doing cosine similarity with the dot product in the case of unit vectors. If you look at the formula, the cosine of the angle between vectors is the dot product divided by the product of the lengths of each vector. So in this case the length of each vector is 1, so your denominator is 1. So all you have in the numerator is the dot product. So the cosine of the angle is the dot product!

[ruby_coder]

More discussion and testing here:

Some questions about text-embedding-ada-002’s embedding

We want to use the embedding generated by the text-embedding-ada-002 model for some search operations in our business, but we encountered a problem when using it. Here are two texts. text1: I need to solve the problem with money text2: Anything you would like to share? following is the code: emb = openai.Embedding.create(input=[text1,text2], engine=model,request_timeout =3) emb1 = np.asarray(emb.data[0]["embedding”]) emb2 = np.asarray(emb.data[1]["embedding”]) def cosine_similarity(a, b): …

[rex.vanhorn]

@raymonddavey is exactly right. It doesn’t have to be a csv file, but not that it cannot be an xlsx file because the Excel cells can’t hold the vector because the vector is too big. CSV files work really well, though.