Danila Sinopalnikov

[Olivier Bachem @OlivierBachem]

Gemma running in space! I don't think we had this on the radar when we trained the model!

[Olivier Bachem @OlivierBachem]

Really proud of everyone in the team that worked hard for this release: Gemma 4 is here! https://t.co/TJI23XAFGL

[Olivier Bachem @OlivierBachem]

Our goal in the Gemma team is to ship models that are useful by generalizing to unseen tasks. Hence, we are extremely strict about not doing anything that would target specific benchmarks instead of teaching the model broad capabilities. It's great to see this reflected here.

[Henry Shevlin @dioscuri]

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.

[Robert Dadashi @robdadashi]

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 !

[Robert Dadashi @robdadashi]

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

[Geoffrey Cideron @CdrGeo]

Happy to introduce our paper MusicRL, the first music generation system finetuned with human preferences.

Paper link: https://t.co/81wIOSDTTa

[Ibrahim Alabdulmohsin | إبراهيم العبدالمحسن @ibomohsin]

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!
https://t.co/h2m9gsisVp

[Google DeepMind @GoogleDeepMind]

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: https://t.co/QreStPR4wd
Quickstart: https://t.co/a6ZWPdx0lj
V2 Paper: https://t.co/rvXJFbwnsD 1/

[Olivier Bachem @OlivierBachem]

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.
https://t.co/q7WPeybZjh

[Johan Ferret @johanferret]

🚨 New preprint! 🚨
"There is no Turning Back: A Self-Supervised Approach to Reversibility-Aware RL"
w/ @NGrinsztajn & others

📜 https://t.co/wlEEVMcFqI

Want to estimate how hard to reverse actions are?
Use reversibility for better exploration and control?
We got you!

🧵👇

[Google AI @GoogleAI]

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 https://t.co/w5cn8p51x1

[Danila Sinopalnikov @sinopalnikov]

I liked a @YouTube video https://t.co/cyUXOLGi70 Two Blades ski club

[Danila Sinopalnikov @sinopalnikov]

@foxnewsdesk Yes, you can use the photo

[Danila Sinopalnikov @sinopalnikov]

#GoogleWalkout in Zürich

[Danila Sinopalnikov @sinopalnikov]

@ZutPetula @letemps Yes

[Danila Sinopalnikov @sinopalnikov]

@NoahChipley Yes

[Danila Sinopalnikov @sinopalnikov]

@flordjones Yes

[Danila Sinopalnikov @sinopalnikov]

@TakakiShigure Yes

[Danila Sinopalnikov @sinopalnikov]

@adamparkersky Yes

[Danila Sinopalnikov @sinopalnikov]

@Hamilbug Yes