🏅 Top traders

GO GO GO GO

Well what's the project

@vluzko, we're doing mechanistic interpretability on models trained on tasks in which grokking occurs, similar to Neel Nanda's analysis of a transformer trained on modular arithmetic. Our goal is to interpret models trained on other tasks (e.g. permutation composition and image brightness classification) where grokking occurs.

you can do it, bet against me!

Will @Kaivu and I's final project for 6.S898 result in an arXiv paper by June 2023?, 8k, beautiful, illustration, trending on art station, picture of the day, epic composition

If the work is part of a larger paper, what happens?

The market will resolve YES.