Talk with Neel Nanda (Google DeepMind)

​Neel Nanda, who leads the mechanistic interpretability team at Google DeepMind, is returning to OAISI for a talk this term!

​The talk’s title is “An overview of interpretability and how it can and cannot help AI safety”. In it, Neel will cover what interpretability research is, why risks from advanced AI matter, and where today's techniques still fall short.

​Last year, OAISI welcomed ~270 people to hear Neel give a deep dive into interpretability and frontier AI research. Those of you who attended will remember it filling the entire lecture theatre plus an overflow room. So, if you missed that, this is your chance.

​No technical background is needed. After the talk, you'll have the chance to chat more with Neel and OAISI members over pizza.

​This talk will not be recorded or livestreamed, so be sure to come along in person!

​Subscribe to our events calendar: https://luma.com/oaisi

​Looking forward to seeing you there :)