Person
Dan Hendrycks
Dan Hendrycks is the executive director of the Center for AI Safety, which he co-founded in 2022, and an adviser to xAI and Scale AI. A UC Berkeley PhD, he devised the GELU activation function and the widely used MMLU benchmark for measuring language-model knowledge.
Appears alongside
Featured in threads
Tracks
- Benchmarks & progress 2
CAIS and Scale AI unveil Humanity's Last Exam results
A 2,500-question expert benchmark built from submissions by nearly 1,000 academics found every frontier model, including o1 and GPT-4o, scored under 10%.
Benchmarks & progress
Hendrycks et al. publish the MMLU benchmark
15,908-question, 57-subject multiple-choice benchmark spanning elementary to professional level; GPT-3 improved on random chance by roughly 20 points on average.
Benchmarks & progress