Beth Barnes leads ARC Evals. She designs ARC’s evaluations of generative AI models and oversees a growing technical team carrying them out. Beth previously worked with DeepMind’s Chief Scientist on scaling laws for forecasting deep learning progress and at OpenAI where she helped develop safety targets and evaluated scalable oversight techniques for alignment and code models for misalignment before release. She has written and presented on a range of technical and theoretical issues related to aligning machine learning systems with human values.
