Towards Monosemanticity: A step towards understanding large language models

Take a look under the hood

How LLMs Think