Accepted to CogSci 2024 and Reinforcement Learning Conference 2024.
How can AI agents learn motivations from others and internalise social rewards?
Member abstract at CogSci 2023.
How can we computationally model infant learning and attachment styles?
Demo paper at EDM 2023.
How do human learners perform on a puzzle video game compared to AI learners?
Accepted to TMLR 2023.
How can we use AI assistants to generate metaphors?
How well do large language models perform on the MATH dataset?
How well do large language models perform on an unnatural in-context learning task?
What are the next steps in teaching foundation models formal reasoning and using them?
Inspired by all the scaling laws excitement, I coded up a numerical simulation of neural networks before GPUs were popular for science and leaderboard points!
© 2019–2025