Research Manager, Interpretability

About the Interpretability team:

When you see what modern language models are capable of, do you wonder, "How do these things work? How can we trust them?"

The Interpretability team’s mission is to reverse engineer how trained models work, and Interpretability research is one of Anthropic’s core research bets on AI safety. We believe that a mechanistic understanding is the most robust way to make advanced systems safe. 

People mean many different things by "interpretability". We're focused on mechanistic interpretability, which aims to discover how neural network parameters map to meaningful algorithms. Some useful analogies might be to think of us as trying to do "biology" or "neuroscience" of neural networks, or as treating neural networks as binary computer programs we're trying to "reverse engineer".

We recently showed that we could extract millions of meaningful features from  Anthropic’s production Claude 3.0 Sonnet model, along with an initial demonstration of how we can use these features to change the model’s behavior by creating “Golden Gate Claude”. Achieving these results required a large engineering effort including optimizing sparse autoencoders (SAEs) across many GPUs, and building tools to visualize millions of features. Work like this is central to our roadmap of using mechanistic interpretability to improve the safety of LLMs like Claude.

A few places to learn more about our work and team are this introduction to Interpretability from our research lead, Chris Olah; a discussion of our work on the Hard Fork podcast produced by the New York Times, and this blog post (and accompanying video) sharing more about some of the engineering challenges we’d had to solve to get these results. Some of our team's notable publications include Scaling Monosemanticity: Extracting Interpretable Features from Claude 3 Sonnet, Towards Monosemanticity: Decomposing Language Models With Dictionary Learning, A Mathematical Framework for Transformer Circuits, In-context Learning and Induction Heads, and Toy Models of Superposition. This work builds on ideas from members' work prior to Anthropic such as the original circuits thread, Multimodal Neurons, Activation Atlases, and Building Blocks.

About the role:

As a manager on the Interpretability team, you'll support a team of expert researchers and engineers who are trying to understand at a deep, mechanistic level, how modern large language models work internally. 

Few things can accelerate this work more than great managers. Your work as manager will be critical in making sure that our fast-growing team is able to meet its ambitious safety research goals over the coming years. In this role, you will partner closely with an individual contributor research lead to drive the team's success, translating cutting-edge research ideas into tangible goals and overseeing their execution. You will manage team execution, careers and performance, facilitate relationships within and across teams, and drive the hiring pipeline. 

If you're more interested in making individual direct technical contributions to our research, feel free to apply to our Research Scientist or Research Engineer roles instead.