Extract similar word from model

Hello! Can I extract a similar word from the model? Also, is it possible to assess the similarity between two words, not sentences, but only words? I’ll hope you help.

There is not nearly enough information in your question to start thinking about how to help you!

Please remember that we don’t know anything about the problem that you are facing. Please provide enough context so that someone who has no clue about what you are dealing with (but has some experience working with ML and PyTorch) can understand the issue.

I agree with @gphilip, we need a bit more information to comprehent your question fully.
However, perhaps the NLTK WordNet helps, to get similar words (synset) as well as compute the similarity between two words.

I want to extract similar words by context from a language model like bert. For example, model.most_similar from gensim by word2vec, also for the transformers language model.