It shouldn't be much of a surprise that a model whose central feature is "finding high-dimensional associations" would be able to identify and semantically grou

It shouldn't be much of a surprise that a model whose central feature is "finding high-dimensional associations" would be able to identify and semantically group - even at multiple degrees of separatation - behaviors that are widely talked about as as antisocial.