🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

Wrapped up Stanford CS336 (Language Models from Scratch), taught with an amazing team Tatsunori Hashimoto Marcel Rød Neil Band Rohith Kuditipudi. Researchers are becoming detached from the technical details of how LMs work. In CS336, we try to fix that by having students build everything:

We are looking for more reviewers. If you are interested, please fill out this form: docs.google.com/forms/d/e/1FAI…

pls review for our workshop 🥺

Another #ICML2025 paper! Why Has Predicting Downstream Capabilities of Frontier AI Models with Scale Remained Elusive? TLDR: Predicting language model performance with scale on multiple choice question-answer (MCQA) benchmarks is made difficult b/c ... 1/3

Third #ICML2025 paper! What effect will web-scale synthetic data have on future deep generative models? Collapse or Thrive? Perils and Promises of Synthetic Data in a Self-Generating World 🔄 Joshua Kazdan Apratim Dey Matthias Gerstgrasser Rafael Rafailov @ NeurIPS Sanmi Koyejo 1/7

Joshua Kazdan Lastly, don't sleep on our NEW position paper: Model Collapse Does Not Mean What You Think We discuss the state of research on synthetic data & model collapse, and where we feel more effort is necessary w/ Alvan Arulandu Joshua Kazdan Sanmi Koyejo arxiv.org/abs/2503.03150 7/7

Really enjoyed this talk by Harry Goldstein that demonstrates inventive uses of the #LeanLang InfoView enhanced by metaprogramming techniques to display real-time testing data. #LeanProver #Metaprogramming #VSCode #PropertyTesting

New position paper! Machine Learning Conferences Should Establish a “Refutations and Critiques” Track Joint w/ Sanmi Koyejo Joshua Kazdan Yegor Denisov-Blanch Francesco Orabona Koustuv Sinha Jessica Zosa Forde Jesse Dodge Susan Zhang Brando Miranda Matthias Gerstgrasser isha Elyas Obbad 1/6

Tuğrulcan Elmas 🇹 🇯 Rylan Schaeffer Sanmi Koyejo Joshua Kazdan Yegor Denisov-Blanch Francesco Orabona Koustuv Sinha Jessica Zosa Forde Jesse Dodge Susan Zhang Matthias Gerstgrasser isha Elyas Obbad One fix: invite the peers to be authors. Being generous & correcting the field is most important. I'd be happy to write refutations for my papers. Discussion & truth matter most. The shift socially should be: being more open to corrections than feeling attacked so easily.

I'm in ICML too! Code and math -- especially ai for formal methods and lean!

I'll be at ICML Conference #ICML2025 next week to present three papers - reach out if you want to chat about generative AI, scaling laws, synthetic data or any other AI topic! #1 How Do Large Language Monkeys Get Their Power (Laws)? x.com/RylanSchaeffer…

#2 Why Has Predicting Downstream Capabilities of Frontier AI Models with Scale Remained Elusive? x.com/RylanSchaeffer…

#3 Collapse or Thrive? Perils and Promises of Synthetic Data in a Self-Generating World 🔄 x.com/RylanSchaeffer…

Interesting!

Same! Chat with me about agents and automatic prompts stuff! So exciting :)

Someone know how to find my old tweprints?

Someone know how to find my old tweprints?

I still find it mysterious whether and how intelligence and capabilities transfer between domains and skills - from meta learning during early days to more recent question like whether solving maths helps writing a good essay. Sometime I feel a bit pessimistic given not enough

Anyone know if this is still happening? Not on website, :( #ICML2025 ICML Conference