Also yin and yang. Models should be aware of hate and anti-social topics and training data. Removing it all in the hopes of creating a "pure" model that can nev

Also yin and yang. Models should be aware of hate and anti-social topics and training data. Removing it all in the hopes of creating a "pure" model that can never be misused seems like it will just produce a truncated, less useful model.