Research
-
Deduction-Projection Estimators for Understanding Neural Networks (2025) [pdf]
Gabriel Wu, advised by Sitan Chen
Undergraduate thesis presented to Department of Computer Science -
Estimating the Probabilities of Rare Outputs in Language Models (2024) [arXiv] [Blog post]
Gabriel Wu, Jacob Hilton
International Conference on Learning Representations (ICLR 2025), Spotlight paper
Work done at the Alignment Research Center in Berkeley, CA -
Harvard Undergraduate Survey on Generative AI (2024) [arXiv]
Shikoh Hirabayashi, Rishab Jain, Nikola Jurkovic, Gabriel Wu
Led the first Harvard survey studying student perspectives on AI -
Testing Tensor Products of Algebraic Codes (2023) [arXiv]
Sumegha Garg, Madhu Sudan, Gabriel Wu