🔥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

As expected, that was popular. Here is my attempt at consolidating all the answers into a list. - Prenorm: normalization in the residual blocks before the attention operation and the FFN respectively - GQA (Group Query Attention): more Q than (K, V)

I'm back and Gemini 2.5 Pro is still the king (no glaze) I did some more manual data cleaning and scrapped the shitty "average scaled score" and replaced it with Glicko-2 rating system with params: INITIAL_RATING = 1500 INITIAL_RD = 350 INITIAL_VOL = 0.06 TAU (τ) =

NeurIPS deadline came early. I remember someone proposing stochastic conference deadlines with exponential decay, is this an A/B test for that?

BREAKING: They now are sending secret service agents to James Comey's house for calling for the removal of a sitting president who is going against the Constitution. If they don't like your "free speech" they will intimidate you, lie about you, and make you the enemy. 8647

i miss gemini-2.5-pro-exp-03-25 so bad :(

In BO, normally we use P(y|x,D) for AF calculation. With ACE, we can get P(xopt|D) (just predict the optimum!), making Thompson sampling and entropy-based AF straightforward to compute. See Luigi's latest blog post on how we craft the dataset and the broader direction beyond BO!

Now accepted at JMLR, and with an extension to general finite groups (including non-abelian groups)! Updated version of our (w/ Stéphane Deny) work: arxiv.org/abs/2412.11521

Claude: my wife and I went antique shopping this weekend Gemini: if I can’t get this code to work I will k*** myself ChatGPT: the answer to your question came to me in a dream Grok: why yes I was in Berlin in 1939, why do you ask?

I didn't want to post on Grok safety since I work at a competitor, but it's not about competition. I appreciate the scientists and engineers at xAI but the way safety was handled is completely irresponsible. Thread below.

So, all the models underperform humans on the new International Mathematical Olympiad questions, and Grok-4 is especially bad on it, even with best-of-n selection? Unbelievable!

Remember to follow the official rebuttal guide.

I have an opening for a 2-year postdoc in probabilistic machine learning and/or experimental design. The application deadline is the 3rd of September. See here for details and how to apply: tinyurl.com/rainmlpostdoc2…