Take a Look Under the hood | Towards Data Science

Using Monosemanticity to understand the concepts a Large Language Model learned