Stephen McAleese comments on An Introduction to Representation Engineering—an activation-based paradigm for controlling LLMs