But when people use the term bias in NLP, that’s what they’re talking about. They don’t want an aspect of the model to do something it ought not do. It’s a case

But when people use the term bias in NLP, that’s what they’re talking about. They don’t want an aspect of the model to do something it ought not do. It’s a case of omitted variable bias causing things like the word analogy issues you hear about. Not an issue of bias in predicting the masked word.