Bertopic cluster question

Nobody has claimed this yet.

Hello,

I'm using BERTtopic for topic modelling and sometimes I find that it clusters topics more broadly than I'd like.

Question 1:Is there a general approach or parameter I can tweak to control the granularity of the topic clustering?

For instance, if I have two sentences: "I like to eat apples" and "I like to eat bananas", I'd prefer them to be classified into two distinct topics given my emphasis on the specific food being liked. Currently, they might be grouped into one topic. How can I adjust the model or the algorithm to achieve this desired granularity in topic classification?

Question2: I found sometimes for the topic -1, outliter there is too many sentence. Is there any way to reduce the noise? Actually I feel some noise actually is misclassify as noise
Thank you!

Start by investigating how BERTopic controls topic granularity and how it assigns documents to topic -1. Determine whether the requested behavior is supported by existing parameters or would require a broader algorithmic change; the work is complete only when both questions have a documented, reproducible resolution.

Written by the indexing model from the issue text.