This research continues many of the directions our team worked on prior to Anthropic, including: GPT-3, Circuit-Based Interpretability, Multimodal Neurons, Scaling Laws, AI & Compute, Concrete Problems in AI Safety, and Learning from Human Preferences. You will work across multiple segments and products, partnering with analytics engineers, fellow data scientists, and go-to-market leadership to turn complex commercial data into actionable strategy.