Technical interview with Neel Nanda

[dmm]

www.lesswrong.com/posts/r2yTwkGt3kbQG2mXi/axrp-episode-19-mechanistic-interpretability-with-neel-nanda

Comments

[dmm]

This is a Feb 3 interview.

Neel Nanda solved the mystery of "Grokking" as an independent researcher.

Before that he co-authored other cool papers about internal mechanisms of Transformer-based models while at Anthropic AI.

[dmm]

See my notes at https://dmm.dreamwidth.org/64571.html

and https://github.com/anhinga/2022-notes/tree/main/Grokking-is-solved

[dmm]

"Hello everybody. In this episode I’ll be speaking with Neel Nanda. After graduating from Cambridge, Neel did a variety of research internships, including one with me, before joining Anthropic where he worked with Chris Olah on the Transformer Circuits agenda. As we record this, he’s pursuing independent research and producing resources to help build the field of mechanistic interpretability. But around when this episode will likely be released, he’ll be joining the language model interpretability team at DeepMind."