WorkThe story, in brief
Gemma Scope 2: helping the AI safety community deepen understanding of complex language model behavior
Google just open-sourced interpretability tools for an entire model family. Here's why the AI safety community is paying attention.
The KeyNews take
Why it matters
Google's Gemma Scope 2 release democratizes model interpretability across the Gemma 3 family, giving the safety and research community open-source tools to audit and understand LLM behavior—a critical capability as models become more complex and deployed at scale.
The key facts
5 to knowGemma Scope 2 released with open interpretability tools
Tools available across entire Gemma 3 model family
Targets AI safety community and model behavior research
Open-source distribution model
Released Dec 16, 2025
Go to the source
Google DeepMind Blogdeepmind.google
Publisher excerpt: Open interpretability tools for language models are now available across the entire Gemma 3 family with the release of Gemma Scope 2.