Daniel M. Ziegler

Measuring AI Ability to Complete Long Tasks

preprint

Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training

arXiv 2024

Fine-Tuning Language Models from Human Preferences

arXiv 2019

Paul Christiano

Head of AI Safety

Adam Jermyn

Alec Radford

researcher

Amanda Askell

researcher

Amy Deng

researcher

Ansh Radhakrishnan

Ben West

researcher

Beth Barnes

founder / CEO

Brian Goodrich

Buck Shlegeris