Mechanistic Interpretability Reading Group
A weekly reading group for anyone interested in mechanistic interpretability research. Our goal is to read papers from the mech interp community, share ideas, and deepen our understanding of how neural networks work under the hood.
Each week, we pick a paper to read and discuss together. The paper for the week will typically be posted on Monday, giving everyone time to read ahead.
Everyone is welcome to present a paper or suggest one for a future session. If you have paper suggestions or want to present, please fill out the forms below. — all levels of familiarity with the field are welcome!
For this week, we'll be reading Towards Automated Circuit Discovery for Mechanistic Interpretability
You can view our full calendar here: https://adityaiyer7.github.io/blogs/reading-group/
If you are interested in presenting a paper for a future session, please fill out the form here: https://forms.gle/wpi5tJbi8CrKikEi8
If you would like to suggest a paper (but not present), please fill out the form here: https://forms.gle/cfK2fPTtWNP8Gw9a6