Hi, I'm Ryan.
I try to understand what is actually going on inside language models, their strange and alien biology, and I work on making them safe. Currently I am an interpretability and alignment researcher at d_model.
I love patterns everywhere, from nature and history to books and math. Here is one of my favorites.
Along the way I've done a bunch of things, including independent mech interp research, creating and teaching a course on transformers, and mentoring junior researchers.
Papers are on Scholar, code is on GitHub. If something here is interesting, feel free to write to me or grab a time!