large distances in plot_embeddings results

[JeremyPohly]

I am trying to use the plot_embeddings function on
scopus_cleaned.csv file. When I do, the distances between the first topic and all the others is so great that they all collapse down to one unreadable point:

Any ideas what's going on here? Can I ignore all of the "1" categories somehow?

Code to reproduce is simply:

docs = litstudy.load_csv('scopus_cleaned.csv')
corpus = litstudy.build_corpus(docs, ngram_threshold=0.5)
num_topics = 20
topic_model = litstudy.train_nmf_model(corpus, num_topics, max_iter=300)
ax = plt.figure(figsize=(6, 6), dpi=150).add_subplot(111)
litstudy.plot_embedding(corpus, topic_model, ax=ax)