Awesome and highly useful: FineWeb-Edu 📚👏 High quality LLM dataset filtering the original 15 trillion FineWeb tokens to 1.3 trillion of the highest (educational) quality, as judged by a Llama 3 70B. +A highly detailed paper. Turns out that LLMs learn a lot better and faster

Guess who's back... Paul Clist Christoph Görtz

It's quite a feat from england being more boring than the most boring country in the world. (Stolen from Jaffer)

I'm going to learn French by reading T'Choupi and listening to nursery rhymes on a road trip. Regarde un Bateau sur l'eau.

The thing our American colleagues can't comprehend is just quite how much better French food is.

I just asked GPT4 to explain me ISO 42001 and it told me it was a Cannabis Management System

📢Our paper on "Understanding Child Sex Trafficking Using Victim-Level Data" is forthcoming in EDCC. The fieldwork for this paper was one of the toughest in my life, and I am glad the paper is out, and the world will learn so we can try to change. journals.uchicago.edu/doi/10.1086/73…

For Lebanon. Stay strong. youtu.be/bSpUITpq_1A?fe…

100% my phd

Phew

This is great, but it had to be in Oxford didn't it? Tom Forth

This is essential research. AI can do what the web cannot and give entire communities new tools to learn and create.

Watching the UK slowly crumble, I’m increasingly convinced its uninformed attitudes like this - that the world is London + Oxbridge, that is behind the slow decline. Hard to get innovation from 13th century spires.

nobody called nobody texted nobody needs you POST-SCARCITY BLUES [Found footage from a Boring Dystopia]>>>SOUND ON what are you for?

the classroom scene at the end.... i do sometimes wonder if we'll regret this new technology wave.

All that money and it scores 56% in early grade visual maths. benchmarks.ai-for-education.org/leaderboards/m…

For the Brits this printed out would be 6.5x the area of Wales. Or 181 million london buses.

Non bon