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. Hands-on engineering experience in a weapons-relevant technical domain, such as systems engineering, robotics and autonomy, guidance/navigation/control, sensors and signal processing, aerospace or mechanical engineering, materials and energetics, or embedded software.