Alexa Pan 潘悦

Hi, I'm Alexa. I'm a Member of Technical Staff at Redwood Research. I currently think about methodology for evaluating alignment and assessing alignment risk.

My research is largely motivated by mitigating worst-case risks from AI. Apart from this, I'm also interested in creating better futures and developing a better philosophy and science of non-human minds.

Before Redwood, I worked at Forecasting Research Institute; before that, I explored research on cooperative AI and LLM lie detection.

I have a BA in Physics and Philosophy from Yale. I wrote my thesis on the rationality of value change, advised by the wonderful L.A. Paul.

I grew up in China and Singapore. Outside of research, I enjoy reading and writing poetry, photography, (literal and figurative) random walks, and live music. I'm always excited to connect with machine learning (safety) researchers and philosophers or anyone curious about my work.