It's been an incredible week! - Proud to be part of the team behind ModernBERT! It's a much needed upgrade on BERT across the board! Excited to see it trending on the HF Hub! 📈 - 1st place in the EEDI @kaggle competition - New milestone: reaching #4 on @kaggle global rankings!

@Dorialexander non-english user prompts from the lmsys-chat-1m dataset should be relevant:

@GriffinAdams92 @AIatMeta Congrats Griffin!! 🎉 Looking forward to what you build at Meta!

@implisci Nice! I did my phd from NUS during 2014-18! After that I've been working in data science related roles.

@Javihm77 @kaggle Here is the blogpost:

@implisci Discussion link:

@dk21 Thanks Darek for the kind words! Would love to see you on kaggle comps soon!

@dk21 @kaggle Thanks Darek! 🙂

@GriffinAdams92 @kaggle Thanks Griffin! Let's catch up soon!

@bclavie @kaggle Thanks Ben! o1-preview was ~0.67, really happy to be within 4 pct points. Now with o3, not sure where is the ceiling!

@Nate_Keating @kaggle Thanks Nate! Yes, absolutely, ModernBERT most likely would be the king of efficiency tracks!

I'm grateful for the learning opportunities and amazing collaborations! Looking forward to more exciting challenges in 2025!

It's been a great learning experience!

RT @bclavie: It's finally out! We at @answerdotai, @LightOnIO and friends are releasing ModernBERT 🎉 It does exactly what it says on the…

RT @ncooper57: As a lead researcher at @stabilityai, I worked a lot with synthetic data to train LLMs and VLMs. It is the most underrated w…