Doing Matrix Calculus can be messy, specially when we need higher order derivatives. Writing them out using Tensor Diagrams makes even the Hessian Chain Rule relatively simple:

Such a fun video!!!

This book is having a moment. I read this book twice in my first year of graduate school because it felt so important for me to understand to be an algebraic geometer. Ultimately it wasn’t important but it was a wonderful effort.

A recent ChatGPT behavior on data tasks is to tell me that it needs some time to complete the task. Note: it never does it

bay area humans saw this daily and decided to invent b2b saas

This is a good post.

I'm excited to share several repos, datasets and models for SAEs on sentence embeddings! First: a visual interface to the features learned by an SAE, inspired by neuronpedia search and browse features, see top activating examples and find similar features links below

Excerpt from a video I just posted on YouTube (about the multilayer perception layers in a transformer, and how LLMs may store facts).

New life goal: generate a shitty magic the gathering combo deck that generates an arbitrarily large number of lands and then ask a judge to rule if it’s a prime number because of this trigger

It’s not that employees need to constantly be watching their bosses but rto policies just improve the work ethic organically /s

There are a number of apps that if I could just download the entire database to my phone, I would. I don’t need connectivity on everything, just the data please.

Inspired by this tweet… Tell me your environmental productivity hacks that use LLMs. How are you making your digital life a little better with a dash of LLM?

All I see is category theory

4 million lichess.org chess puzzles

For some reason my watch and headphones got desynced and I stopped getting route directions. Ok nbd. *several hours later* Connect my headphones to my phone and turn on a podcast, get about 25 route directions in a row for a hike I finished this morning. 😅

a collection of set theorists who are so excited that they cannot contain themselves