I work as an ML researcher, currently interested in language models (with an increasing focus on their internals, e.g., representations, model merging, etc.). In the past I have also worked on Bayesian modelling, causal inference, and information theory applications to the previous.

I'm also slowly working through the basics of category theory as well as persistent homology. I'd love to understand how ideas from cat. theory map to applied questions, e.g., the work of Fong, Spivak, et al., with Bayesian networks and data manipulations. For persistent homology, I want to see how tools can be used to learn more about how language models tick.