Grateful for chatting with Sam Charrington about LLM reasoning, test-time scaling & s1!

Had a great time giving a talk on s1 at Microsoft GenAI! I enjoy talks most when they're not a monologue but rather a back-and-forth with new ideas that go beyond the paper. This was one of those thanks to an amazing audience with hard questions😅 youtube.com/watch?v=EEkxuq…

Finetuning on raw DeepSeek R1 reasoning traces makes models overthink. One of our early s1 versions was overthinking so much, it questioned the purpose of math when just asking what's 1+1😁 Retro-Search by Ximing Lu & team reduces overthinking + improves performance!

Reasoning & test-time scaling don't just matter for generating text with LLMs — Rulin Shao, Rui Qiao & team show how these are key to retrieval quality. ReasonIR is SoTA on reasoning-intensive retrieval across multiple test-time compute budgets!

Very excited to join KnightHennessy scholars at Stanford🌲 Loved discussing the big goals other scholars are after — from driving Moore’s Law in biotech to preserving culture via 3D imaging. Personally, most excited about AI that can one day help us cure all diseases :)

Interesting seeing what are the current best personal assistants that live only in your terminal! Follow Mike A. Merrill Alex Shaw & team for future updates on Terminal-Bench :)

Nice work by Ryan Marten Etash Guha & co! Made me wonder --- if you aim to train the best 7B model where there are much better (but much larger) models available, when does it make sense to do RL over distill+sft?🤔

Congrats to Matt Deitke, Chris, Ani Kembhavi & team for the Molmo Award!! Fond memories of us hurrying to fix model inference until just before the Sep release😁

Scaling Data-Constrained LMs is now also in JMLR: jmlr.org/papers/v26/24-… Looking back at it 2yrs later, repeating & mixing seem standard now, but maybe another powerful lever to scale data-constrained LMs turns out to have been RL - arguably underrated back then!