Engage personally with C-level and senior technical stakeholders (CIOs, CTOs, Chief Actuaries, Heads of Data/AI) - building business cases and value narratives, and navigating complex procurement, security, regulatory, and legal review through to production deployment and expansion. 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.