How to reduce outliers after having performed hierarchical topic modeling?

[HastyBeaver]

Hi everybody,

After finding topics and probabilities using the following code:
topics, probs = topic_model.fit_transform(train_documents, embeddings)
I wanted to perform hierarchical topic reduction. After visualising the hierarchy, I decided to merge a few topics together using the merge_topics function. I also visualized the pd.Dataframe printed by get_topic_info, and, obviously, I can see the new topics as well as the new outliers.
Now, I want to apply reduce_outliers, but it seems that this function requires the topics list as a parameter, which is supposedly the topic list after the hierarchical reduction.
How can I obtain this? I think I may be missing something, but I can't seem to find a solution.

Thank you!

[MaartenGr]

You can access any updated topics with topioc_model.topics_ and use those for further calculations/processes.