AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
2min chapter
#122 - AI for Word and Excel, leaked Google memo, ImageBind, LLAva, robot soccer, Midjourney 5.1
Last Week in AI
How to Find Bad Neurons That Do Things You Don't Want
This is a pretty different approach from what you've seen, even if it is pretty, I don't know, logical in terms of how it is implemented. It's not surprising that this works. You can do this as a human pretty easily. Now, the other question is, how is this actually useful beyond equinability? There's an argument to be made that you could find bad neurons that do things you don't want. Maybe you can edit out certain knowledge or certain ideas within a language model. So some caveats on the impacts of this and sort of the significance, but still it is very cool.
00:00
Transcript
Episode notes
Get the Snipd
podcast app
Unlock the knowledge in podcasts with the podcast player of the future.
AI-powered
Discover
highlights
Listen to the best highlights from the podcasts you love and dive into the full episode
Save any
moment
Hear something you like? Tap your headphones to save it with AI-generated key takeaways
Share
& Export
Send highlights to Twitter, WhatsApp or export them to Notion, Readwise & more
AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
Discover
highlights
Listen to the best highlights from the podcasts you love and dive into the full episode