Gemma Scope: helping the safety community shed light on the inner workings of language models
Technologies Published 31 July 2024 Authors Language Model Interpretability team Announcing a comprehensive, open suite of sparse autoencoders for language ...
Read more
