Gabriel Mukobi comments on Addendum: basic facts about language models during training