I typically spend (perhaps) too much time prepping for talks I give - brainstorm different ways to communicate the message. Folks who stopped me to thank and folks who bothered to email me to thank me - you are all the reasons why this is worth it. Thank you♥️

Tuesday 10:45am-12:45pm
(Spotlight) #1526: Does Localization Inform Editing? Surprising Differences in Causality-Based Localization vs. Knowledge Editing in Language Models
With Peter Hase, Mohit Bansal , and Been Kim

Last but not least - our work at #NeurIPS2023 on creating embedding space for RL agents to act and explain.

Our work on questioning the effectiveness of localization methods for editing at #NeurIPS2023 !

Our work on probing NN with ideas from Gaussian process at #NeurIPS2023 !

Hope to convince folks who are new in ML to 'wing it' on Monday at Neurips.. See you there. :)

This video is worth watching. :)

This work would have not been possible without Lisa Schut - a previous professional chess player and herself a champion, and now a PhD student in ML at Oxford! One of those once in a lifetime kind of collaboration ♥️

Very excited to talk about Leveraging AlphaZero to Improve our Understanding & Creativity in Chess ♟️🤯 with Been Kim at the Stanford HAI Fall Conference!
In this work, we dig into finding chess concepts that are beyond the current collective human knowledge 🧵1/3