Dan Hendrycks
Executive and Research Director of the Center for AI Safety; co-author of MMLU and Humanity's Last Exam
- Executive & Research Director, Center for AI Safety (CAIS)21
Dan Hendrycks directs the Center for AI Safety (CAIS); the organization's site lists Hendrycks as Executive & Research Director. Hendrycks received a PhD from UC Berkeley advised by Dawn Song and Jacob Steinhardt. Hendrycks is first author of the GELU activation function paper and of "Measuring Massive Multitask Language Understanding" (MMLU), and the last-listed author of Humanity's Last Exam, a multi-modal benchmark of expert-written questions across dozens of subjects.12543
Work
Humanity's Last Exam (HLE)Published research
Last-listed author of the paper introducing Humanity's Last Exam, a multi-modal benchmark at the frontier of human knowledge designed as a closed-ended academic benchmark with broad subject coverage.3
Measuring Massive Multitask Language Understanding (MMLU) (external site: arxiv.org)Published research
First author of the 2020 paper proposing a test of a text model's multitask accuracy across 57 tasks, including elementary mathematics, US history, computer science, and law.4
Gaussian Error Linear Units (GELUs) (external site: arxiv.org)Published research
First author (with Kevin Gimpel) of the 2016 paper proposing the GELU activation function, which weights inputs by their value rather than gating them by their sign.51
This profile covers public professional work only, from the sources below. It is not an endorsement or a ranking. To correct or remove it, use the correction form.