Gabriel Dulac-Arnold (@gabepsilon) 's Twitter Profile
Gabriel Dulac-Arnold

@gabepsilon

Gemini Thinking & RL. Poking at neural networks to make them do something. 🔥🌍🔥 🏳️‍🌈 ✊

ID: 14326917

linkhttp://gabe.squirrelsoup.net calendar_today07-04-2008 21:01:36

462 Tweet

765 Followers

594 Following

WikiLeaks (@wikileaks) 's Twitter Profile Photo

JULIAN ASSANGE IS FREE Julian Assange is free. He left Belmarsh maximum security prison on the morning of 24 June, after having spent 1901 days there. He was granted bail by the High Court in London and was released at Stansted airport during the afternoon, where he boarded a

Thomas Mesnard (@mesnard_thomas) 's Twitter Profile Photo

Et voilà Gemma 2! Small models are getting soooo good! Even stronger 2B/9B + our new powerful 27B! Once again, open models by Google DeepMind with state-of-the art pre-trained and fine-tuned checkpoints. Play with it here: aistudio.google.com/app/prompts/ne…

Et voilà Gemma 2! Small models are getting soooo good!
Even stronger 2B/9B + our new powerful 27B!
Once again, open models by <a href="/GoogleDeepMind/">Google DeepMind</a> with state-of-the art pre-trained and fine-tuned checkpoints. 
Play with it here: aistudio.google.com/app/prompts/ne…
Brooklyn Nets France (@brooklynnetsfra) 's Twitter Profile Photo

Je me permets de relayer ce thread de 6 (contre) ARGUMENTS qu'on entend souvent : 1. « Les extrêmes » 2. « Jean-Luc » 3. « Les casseroles de LFI » 4. « Le RN anti-système » 5. « Le RN, on n’a jamais essayé » 6. « Programme irréaliste » Source et crédits à la fin 🔴🟣🔵

Je me permets de relayer ce thread de 6 (contre) ARGUMENTS qu'on entend souvent :
1. « Les extrêmes »
2. « Jean-Luc »
3. « Les casseroles de LFI »
4. « Le RN anti-système »
5. « Le RN, on n’a jamais essayé »
6. « Programme irréaliste »

Source et crédits à la fin 🔴🟣🔵
David Pfau (@pfau) 's Twitter Profile Photo

Biden could literally be in a coma and the whole administration running on autopilot and it would still be better than Trump. I hate that this is where we are at, but it is where we are at.

Marine Tondelier (@marinetondelier) 's Twitter Profile Photo

L'écologie c'est pas punitif. L'écologie c'est plus belle la vie ! Car les première victimes du manque d'écologie sont les plus pauvres, avec le #NouveauFrontPopulaire nous proposons : 👉 mettre le paquet sur la rénovation thermique 👉 rendre les premiers m3 d'eau gratuits 👉

Hugo Laurençon (@hugolaurencon) 's Twitter Profile Photo

Whether you are: •A complete beginner looking to get a high-level overview of the SOTA VLM approaches and their limitations •An expert searching for new directions in the field Our new paper might be worth a read!

Whether you are:

•A complete beginner looking to get a high-level overview of the SOTA VLM approaches and their limitations
•An expert searching for new directions in the field

Our new paper might be worth a read!
Les Soulèvements de la terre (@lessoulevements) 's Twitter Profile Photo

Le 5 octobre : ⛵ Levons les voiles contre l'empire #Bolloré dans l'archipel des Glénan 🧵⤵️ A l'appel du collectif "Lever les voiles" et d'organisations locales et nationales : desarmerbollore.net/agissons 1/12

Le 5 octobre : ⛵ Levons les voiles contre l'empire #Bolloré dans l'archipel des Glénan 🧵⤵️

A l'appel du collectif  "Lever les voiles" et d'organisations locales et nationales : desarmerbollore.net/agissons
1/12
Sven Gowal (@sgowal) 's Twitter Profile Photo

My team at Google DeepMind is hiring. If you are passionate about robust ML, the provenance of synthetic media, and the trustworthiness of data, consider applying: boards.greenhouse.io/deepmind/jobs/…

Pierre Sermanet (@psermanet) 's Twitter Profile Photo

Q: How can we ensure robots behave properly at scale? A: Robot constitutions 📜! Q: How do we verify behavior in undesirable situations at scale? A: Generation! We release the ASIMOV Benchmark for Semantic Safety of robots at asimov-benchmark.github.io Google DeepMind

Q: How can we ensure robots behave properly at scale? A: Robot constitutions 📜!

Q: How do we verify behavior in undesirable situations at scale? A: Generation!

We release the ASIMOV Benchmark for Semantic Safety of robots at asimov-benchmark.github.io

<a href="/GoogleDeepMind/">Google DeepMind</a>
Jeff Dean (@jeffdean) 's Twitter Profile Photo

🥁Introducing Gemini 2.5, our most intelligent model with impressive capabilities in advanced reasoning and coding. Now integrating thinking capabilities, 2.5 Pro Experimental is our most performant Gemini model yet. It’s #1 on lmarena.ai (formerly lmsys.org) leaderboard. 🥇

🥁Introducing Gemini 2.5, our most intelligent model with impressive capabilities in advanced reasoning and coding.

Now integrating thinking capabilities, 2.5 Pro Experimental is our most performant Gemini model yet. It’s #1 on <a href="/lmarena_ai/">lmarena.ai (formerly lmsys.org)</a> leaderboard. 🥇
Noam Shazeer (@noamshazeer) 's Twitter Profile Photo

This model’s “thinking” capabilities are driving major gains: 🧑‍🔬Top performance on math and science benchmarks (AIME, GPQA) 💻Exceptional coding performance (LiveCodeBench) 📈Impressive performance on complex prompts (Humanity’s Last Exam) #1 on lmarena.ai (formerly lmsys.org) leaderboard 🏆