LLM Psychology is a new alignment research agenda, focused on studying LLM systems with a behavioral approach.

This sequence’ goal isn’t to put rigorous framings and build foundations for later research, but to shine a light on this new approach, present potential research avenues, and spark a wave of fresh excitement.

Most of this research has been conducted during SERI Mats 3. I’d like to thank all the people that made this research possible, and @NicholasKees for his mentoring and invaluable advice.

Also, a particular thanks to all the reviewers who helped me make this sequence go from a random 45-pages babble to a (hopefully) interesting sequence of posts:

@Pierre Peigné @Kay Kozaronek @Charbel-Raphaël @NicholasKees