Polysemanticity w/ Dr. Darryl Wright

Jan 22, 2024·
Into AI Safety
Into AI Safety
· 1 min read
Darryl and I discuss his background, how he became interested in machine learning, and a project we are currently working on investigating the penalization of polysemanticity during the training of neural networks.
Polysemanticity Research Diagram, Jacob Haimes
Image by Jacob Haimes / © Kairos.fm / Polysemanticity Research Diagram / CC-BY 4.0

Chapters

01:46 ❙ Interview begins
02:14 ❙ Supernovae classification
08:58 ❙ Penalizing polysemanticity
20:58 ❙ Our "toy model"
30:06 ❙ Task description
32:47 ❙ Addressing hurdles
39:20 ❙ Lessons learned

Links to all articles/papers which are mentioned throughout the episode can be found below, in order of their appearance.

Interview Research
Into AI Safety
Authors
Into AI Safety
One way to make sense of all this tech stuff.

Related