Sublime

https://t.co/Rb6oQBS0ko

It's new post day 🎉 Could a neuroscientist understand an ANN? I cover recent work in mechanistic interpretability at @AnthropicAI led by @ch402 and at UCSB with @naturecomputes @ninamiolane @KlindtDavid https://t.co/6J4wEdRUG6

Learn about updates on spatially-embedded RNNs at @CogCompNeuro: - Implementation in Spiking Networks by Andrew Ham (Poster 3B) - Investigation of structure with entropy measure by @SheeranCornelia (Poster 1A) @DanAkarca @DuncanAstle Summarised:... See more

𝐖𝐡𝐚𝐭 𝐰𝐞 𝐚𝐫𝐞 𝐫𝐞𝐚𝐝𝐢𝐧𝐠 “An integrated single-nucleus and spatial transcriptomics atlas reveals the molecular landscape of the human hippocampus,” by Jacqueline R. Thompson 𝑒𝑡 𝑎𝑙. in 𝑁𝑎𝑡𝑢𝑟𝑒 𝑁𝑒𝑢𝑟𝑜𝑠𝑐𝑖𝑒𝑛𝑐𝑒 https://t.co/52qZb13Dyl

A Swiss lab run by @TrackingActions is collecting data from mouse brains and using it to inform AI systems. More on Hello World with @AshleeVance: https://t.co/RjCm2BGcCM https://t.co/ydXI8AI1U3

A discussion with Iain McGilchrist and Hyman Schipper on Mind, Matter & Metaphysics for the 'Great Minds' series: https://t.co/xuPfP11YFq

The fact that most individual neurons are uninterpretable presents a serious roadblock to a mechanistic understanding of language models. We demonstrate a method for decomposing groups of neurons into interpretable features with the potential to move past that roadblock.

New open-source tool helps to detangle the brain The software tool NeuroTrALE is designed to quickly and efficiently process large amounts of brain imaging data semi-automatically. https://t.co/9zoWXEFs2B

@anilkseth @packyM @cunningstudent @GlobeIdeas @kevin2kelly @erikphoel Another interesting paper is https://t.co/XxUd5YNFky by @davideagleman and @DrDonVaughn