Hobgoblin of Stillwater comments on Revealing Intentionality In Language Models Through AdaVAE Guided Sampling