aogara comments on Hoodwinked: Evaluating Deception Capabilities in Large Language Models