3Blue1Brown is probably my all-time favorite YouTuber. Over the past few months, he released a mini-series on LLMs, diving deep into the attention mechanism, multi-layer perceptrons, and how facts might be stored within the layers of a large language model. 

The visualizations are incredible. Definitely a must-watch if matrix multiplications don’t scare you 😄

Co-Founder & CEO @almanax | AI-powered web3 security | Ex Head of Product @ AnChain AI | UC Berkeley engineering
📍 NYC

Link to the last video of the series: https://youtu.be/9-Jl0dxWQs8