Hi! I’m Haoxing. I am a proud effective altruist, and I’m passionate about helping make sure that the artificial intelligence we develop is safe and beneficial. Currently, I work on evaluating large language models for dangerous capabilities at ARC Evals. Before this I did some research on neural network interpretability at Redwood Research. Check out the work I did on interpreting Leela Zero, a superhuman Go-playing neural network!
Before getting involved in AI safety I spent some years thinking about physics. At Berkeley, I did research in the intersection of machine learning and physics in the amazing Nachman group. Prior to Berkeley, I was fortunate to be part of the Perimeter Scholars International program, and studied at Harvey Mudd College.
When I’m not thinking about technical topics, I like to read, play video games, make music, dabble in agility with my dog Hope, and fly my paraglider.
In a past life, I thought a lot about the foundations of quantum theory. Although I’m no longer pursuing a career in quantum foundations, I remain passionate about the field, and think that many important, recent ideas in quantum foundations are not being effectively communicated to a wider audience. I hope the essays below help bridge that gap just a little. I also hope that my younger self from undergraduate years would have been inspired if she were to stumble upon these.