Applying Representation Model on Non-Leaf Nodes in Hierarchical Topics with BERTopic

Question

I am currently using BERTopic for topic modeling on a set of documents and have integrated an OpenAI model as the representation layer, as outlined in the BERTopic documentation. I am also interested in hierarchical topic modeling, which is detailed here.

In the process of hierarchical modeling, BERTopic provides descriptive labels for the non-leaf topics using the representation model. My question is whether there is a possibility within BERTopic to apply the representation model explicitly to these non-leaf nodes in the hierarchical topics? Specifically, I am looking to generate or refine labels for these higher-level topics based on their aggregated content.

Is there a built-in method to achieve this, or would it require a custom implementation? Any pointers or examples would be greatly appreciated.

I have applied hierarchical topic modeling using the hierarchical_topics method of the fitted topic model. I examined the options available in the hierarchical_topics method to see if there was a way to specify a representation model for non-leaf nodes. However, I could not find any options that would allow me to explicitly apply or adjust the representation model for these nodes within the hierarchical structure.