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. The primitives are the building blocks the rest of Anthropic uses to manage access to models and capabilities: access programs, accounts, organization linking, verifications, user APIs for platforms, and the internal tooling teams use.