Dan Hendrycks
Bio
Dan Hendrycks is the Executive and Research Director of the Center for AI Safety (CAIS), a nonprofit research organization based in San Francisco focused on reducing societal-scale risks from artificial intelligence. He received a B.S. from the University of Chicago in 2018 and a Ph.D. in Computer Science from UC Berkeley in 2022, advised by Dawn Song and Jacob Steinhardt. His research spans machine learning safety, robustness, out-of-distribution detection, and AI ethics. He is the primary author of the GELU activation function (2016), which is widely used in state-of-the-art models including BERT and GPT, and created the MMLU (Massive Multitask Language Understanding) benchmark (2020), one of the most widely used LLM evaluation benchmarks. He also co-developed the MATH benchmark, Humanity's Last Exam (HLE), and authored the 2024 textbook Introduction to AI Safety, Ethics, and Society. He serves as a safety advisor to xAI and Scale AI, both at nominal compensation, and has received early-career funding from EA-aligned organizations for his work on value learning and AI alignment benchmarks.
Links
- Personal Website
- https://danhendrycks.com/
- Twitter / X
- LessWrong
- dan-h
Grants
from Long-Term Future Fund
Discussion
Sign in to join the discussion.
No comments yet. Be the first to share your thoughts.
Details
- Last Updated
- Mar 22, 2026, 3:11 PM UTC
- Created
- Mar 20, 2026, 2:49 AM UTC