Find answers from the community

Updated 3 months ago

Random question... I see qdrant supports

Random question... I see qdrant supports filtering the results by score, and langchain supports returning a score for the qdrant results which is useful. I searched the llama_index docs and didn't see anything similar. Was curios if someone knew if this was something that could be, or was implemented.

5 comments

TTeemu

There is the SimilarityPostprocessor which allows filtering by similarity score: https://docs.llamaindex.ai/en/stable/module_guides/querying/node_postprocessors/node_postprocessors.html#similaritypostprocessor

iicsy7867

Awesome thanks!

iicsy7867

@Teemu I have been reading about this, but I'm having trouble finding more information. Is the range 0-1? Is it the same across different vector stores?

iicsy7867

Also reading up on cosine similarities
https://en.m.wikipedia.org/wiki/Cosine_similarity

So maybe the bounds are [-1,1]? But for this applications I don't think negative values make sense.

TTeemu

It depends on the model you're using, you kind of have to adjust based on that

Add a reply