Google DeepMind — Description: Gemma Scope (2024): 400+ openly released sparse autoencoders for Gemma 2 9B and Gemma 2 2B models, intended to support external interpretability research

NoteThe source directly confirms all key elements of the claim: (1) Gemma Scope is a Google DeepMind project announced in 2024 (July 31, 2024); (2) it contains 400+ sparse autoencoders ('more than 400 sparse autoencoders'); (3) they are for Gemma 2 9B and Gemma 2 2B models; (4) they are openly released and freely available; (5) intended to support interpretability research; (6) Neel Nanda is listed as a core team member. The 'as of 2024-08' in the claim aligns with the July 31, 2024 publication date. The source confirms this was the largest open-source SAE release at the time (prior research 'mainly focused on investigating the inner workings of tiny models or a single layer in larger models'). All claim elements are verified.