About the generation of similar words

[quimqu]

Could you explain what you mean by "context"? Sorry if I missed it.

Context means the environment in which a token appears. It is not about meaning, but about surroundings. It includes which tokens tend to appear nearby, how similar they are, how dense or repetitive the local area is, and also where the token sits in the structure, like its position in the line or the section of the manuscript.