3 ms·
It gets embeddings from open AI and compares them using cosine similarity. Stories are weighted by similarity, dissimilarity and HN score.
by mahin 2y ago
It gets embeddings from open AI and compares them using cosine similarity. Stories are weighted by similarity, dissimilarity and HN score.
- machiaweliczny 2y agoHow you compare a liked set to single embedding? By best match? Or is there some fancier way?
- mahin 2y agoBy averaging the embeddings.