Our work “Do LLMs ``know’’ internally when they follow instructions?” has been accpeted to ICLR 2025!