i think they don't. given enough data those related forms will have similar colocation profiles (aka fingerprints). not a model of semantics.. at scale it might become a proxy though.
Our training dataset is composed of 12 different languages. The English corpus has a size of around 25TB -- ~18TB after a rather aggressive filtering process (char frequencies [i.e. https://gist.github.com/19h/b983769b7ec9c4a9528377d7819892a7], too many unusual features in the sentences, too short sentences, etc.).
Most common misspellings of words have nearly identical fingerprints, and we align the retinas across languages so that translated sentences result in similar fingerprints (via competitive learning).