I think it's a pretty good name for the phenomenon -- maybe the only problem with the term is that what models are doing is 100% hallucination all the time -- i

I think it's a pretty good name for the phenomenon -- maybe the only problem with the term is that what models are doing is 100% hallucination all the time -- it's just that when the hallucinations are useful we don't call them hallucinations -- so maybe that is a problem with the term (not sure if that's what you are getting at). But there's nothing at all different about what the model is doing between these cases -- the models are hallucinating all the time and have no ability to assess when they are hallucinating "right" or "wrong" or useful/non-useful output in any meaningful way.