Can't get updated topic_embeddings_ when using pre-calculated document embeddings?

Nobody has claimed this yet.

### Have you searched existing issues? 🔎

- [x] I have searched and found no existing issues

### Desribe the bug

I can use the `update_topics` function by passing in the appropriate documents and my manually generated topic labels, and using the auto-generated embeddings. This works as intended and I get the updated `topic_embeddings_`.
`topic_model.update_topics(docs, topics=topics)`

However, if I calculate the document embeddings myself and pass them to BERTopic like so:
`topic_model.fit_transform(docs, embeddings=embeddings)`
then the `topic_embeddings_` are not updated after using `update_topics`. This is because the call to `_create_topic_vectors` within `update_topics` is done without any arguments being passed, and would normally use `self.embedding_model`, which is None when I define my own embeddings.

How should I perform the update to get new `topic_embeddings_` in this case?
Could a kwarg for passing pre-calculated embeddings to `update_topics` be added?

### BERTopic Version

Version: 0.16.4

Start by inspecting update_topics and its call to _create_topic_vectors, then reproduce the behavior using fit_transform with pre-calculated embeddings. Trace how the embedding model is stored and used during the update. Done means the documented pre-calculated-embedding workflow updates topic_embeddings_ consistently, with coverage for this case.

Written by the indexing model from the issue text.