Hey everyone,
Has anybody worked with clustering when it comes to embeddings. Essentially grouping embeddings based on similarity, to then be able to, for example, get one record from each group.
What I'm trying to achieve is to be able to create a summary on really long docs. It's way to much to pass the whole text, so what I'm trying to do is to get a chunks from each group and from those create a summary.
Thanks you in advance,
Isaac