Work

I’m currently on sabbatical, and am starting to explore some early stage ideas in both applied interpretability and foundation datasets.

I was at DeepMind from 2019 to late 2023, where I worked on:

• Interpretability for LLMs (e.g. the Hydra Effect, Copy Suppression) and AlphaZero.
• Science of training data.
• RLHF data quality and self-annotation.
• Evaluation of generalist deep RL agents.

I did my PhD (thesis) at Imperial College with Nick Jones and Kevin Murphy.

Research

My papers are listed on my Google Scholar page. I have a list of research projects I’m interested in working on.

If there’s something on there you’re interested in collaborating on, please get in touch!

Writing

I have a substack if you prefer to read there.

• Safety as a scientific pursuit

Contact me

Email is probably best, but you can reach me on Twitter or LinkedIn as well.