Goodfire

Mechanistic interpretability research lab focused on understanding and auditing the internals of large language models