lit VS scattertext

I'd also suggest looking at your data sample, and how your model handles it, with some kind of exploratory analysis tool. Google's Language Interpretability Tool might work for your scenario. This can give you a lot of ideas about how to prepare the data better.

Visualize and understand NLP models with the Language Interpretability Tool The Language Interpretability Tool (LIT) is for researchers and practitioners looking to understand NLP model behavior through a visual, interactive, and extensible tool. Use LIT to ask and answer questions like: What kind of examples does my model perform poorly on? Why did my model make this prediction? Can it attribute it to adversarial behavior, or undesirable priors from the training set? Does my model behave consistently if I change things like textual style, verb tense, or pronoun gender? LIT contains many built-in capabilities but is also customizable, with the ability to add custom interpretability techniques, metrics calculations, counterfactual generators, visualizations, and more.

Use LIT https://github.com/PAIR-code/lit

If what you want is to determine how similar two categories are, or to learn something about the structure or words that compose those categories, you might consider word shift graphs or Scattertext.

Other suggestions: Take a look at Scattertext. Compare keywords to the problem of aspect extraction. I think an underutilized way to look at textual data when you have a single group of interest is the word-frequency-based odds ratio.