New paper: Finetuning on narrow domains leaves traces behind. By looking at the difference in activations before and after finetuning, we can interpret what it was finetuned for. And so can our interpretability agent! 🧵

She shall know your ways as if born to them

I've been really noticing autumn this year

The HS2 bat tunnel is even worse value for money once you factor in the updated direct cash transfer effectiveness estimates.

deleting dating apps because i want to meet someone the old fashioned way (we caught a wild pig together while not sharing a common language, then met 12 years later under the tree we planted)

FYI it's Draft Amnesty Week on TPOB, where users can publish scrappy, draft-y, or incomplete posts with impunity. #DAW

On new beginnings: This week I handed in my notice, ending 10 years in Product Management, capital markets to start as an Alignment Research Manager in January! 🎉