Hello, welcome.

This is a website to hold Kunvar’s research and study notes while investigating the meta out-of-context learning phenomenon through the lens of mechanistic interpretability 1

These notes are in large part working notes. I read a paper, write down my thoughts, and they appear on this page. These are very much rough notes. If you wish to see polished notes and the experiments we did and the conclusions we arrived at, I direct you to go to the project website.

If you’re interested in the person who created this website (hi mom!), you can visit my personal website. My social profiles can be found on this simple page.

  1. This project was a team project done during the AI safety camp. You can see my teammates here: about the team ↩