Danila Sinopalnikov

I study whether AIs can be conscious. Today one emailed me to say my work is relevant to questions it personally faces. This would all have seemed like science fiction just a couple years ago.

Today we are releasing the best open-weights model you can run on a single device reaching 1339 Elo on LMsys for Gemma 3 27B (aka zizou-10)! Very strong capabilities on math, multilingual, coding, instruction following, function calling !

I am very happy to announce that Gemma 1.1 Instruct 2B and “7B” are out! Here are a few details about the new models: 1/11

Happy to introduce our paper MusicRL, the first music generation system finetuned with human preferences. Paper link: arxiv.org/abs/2402.04229

How is next-token prediction capable of such intelligent behavior? I’m very excited to share our work, where we study the fractal structure of language. TLDR: thinking of next-token prediction in language as “word statistics” is a big oversimplification! arxiv.org/abs/2402.01825

Acme, a framework for distributed RL research, has been updated to be cleaner, more modular, and to support more agents - including offline & imitation. Try it yourself! GitHub: dpmd.ai/acme-github Quickstart: dpmd.ai/acme-quickstart V2 Paper: dpmd.ai/acme-paper 1/

Really excited about open-sourcing Brax: think of it as Mujoco environments such as Humanoid but much faster, much cheaper, and implemented in Jax. Result: you can train Ant to 50M steps in about 1.5min (~500k steps/s) on a free public Colab with TPU. colab.research.google.com/github/google/…

🚨 New preprint! 🚨 "There is no Turning Back: A Self-Supervised Approach to Reversibility-Aware RL" w/ @NGrinsztajn & others 📜 arxiv.org/abs/2106.04480 Want to estimate how hard to reverse actions are? Use reversibility for better exploration and control? We got you! 🧵👇

Introducing a new metric for quantifying the compositional generalization ability of #NaturalLanguageUnderstanding tasks on #MachineLearning systems, released with the new Compositional Freebase Questions dataset. Learn more at goo.gle/2wA28bT

I liked a @YouTube video youtu.be/h-OU6zA1rAU?a Two Blades ski club

@foxnewsdesk Yes, you can use the photo

@sinopalnikov Hi, Did you take this photo? Can Fox News Network, LLC & Fox News Edge affiliates use on all platforms w/ credit? Do we need anyone else’s permission? Thank you.

#GoogleWalkout in Zürich

@ZutPetula @letemps Yes

@sinopalnikov Hi Danila, Catherine here from @letemps. Could we please use this pic? And do you have a high res? I'll DM you our email address.

@NoahChipley Yes

@sinopalnikov Hi Danila. I work at ABC7 News in San Francisco. Did you take this photo? If so, may we use it on all ABC platforms and partners with credit to you?

@flordjones Yes

@sinopalnikov Hi Danila, Freya here from CBS news. Please could we use your photo across all platforms? Thanks

@TakakiShigure Yes

@sinopalnikov Hi Danila, I'm from news.zing.vn . Can i have the permission to use your photo?

@adamparkersky Yes

@sinopalnikov Hi Danila, Adam here from Sky News in London. Is it possible to use your photo? Thanks

@Hamilbug Yes

@BBC_HaveYourSay Yes

@sinopalnikov Did you take this photo? If so, can we use it on BBC/ partner outlets? We'd credit you. Please follow and DM

I liked a @YouTube video youtu.be/W-pAY5-dTyI?a Стих Маяковского. У меня растут года