-
Tweets369
-
Followers78
-
Following4K
-
Likes912
Some sexy curves for today, DPO+NLL from IRPO paper:
Microsoft just released Phi-3 - phi-3-mini: 3.8B model trained on 3.3T tokens rivals Mixtral 8x7B and GPT-3.5 - phi-3-medium: 14B model trained on 4.8T tokens w/ 78% on MMLU and 8.9 on MT-bench arxiv.org/abs/2404.14219
We have just released 🍷 FineWeb: 15 trillion tokens of high quality web data. We filtered and deduplicated all CommonCrawl between 2013 and 2024. Models trained on FineWeb outperform RefinedWeb, C4, DolmaV1.6, The Pile and SlimPajama!
HISTORIC MOMENT! Llama-3 70B numbers are INSANE! At 82.0 MMLU, it's FAR AND AWAY the best OSS Model. GSM-8K, Math, and Human Eval are MIND BLOWING as well. The OSS community is definitely going to beat GPT-4 in a matter of weeks!! Xmas came very very early
Announcing OpenChatML. Dolphin will follow this from now on. Go ahead, reply with the xkcd comic, I know you're gonna. But then read the thing and tell me what you think, love your ideas and get something good out there. PRs and discussions welcome.
Wake up honey, new RWKV paper just dropped 🧵⤵️ Paper: arxiv.org/abs/2404.05892 Code: github.com/BlinkDL/RWKV-LM Models: huggingface.co/RWKV (Apache 2.0 license) (1/6)
Buried the lede here a bit, multilingual Mixtral-level performance without all that VRAM needed. Use Megablocks to roll your own DBRX for even better performance. The new 1.6b model outperforms Gemma 2.5b & other edge models handily too, great work by @rikelhood & team.
Buried the lede here a bit, multilingual Mixtral-level performance without all that VRAM needed. Use Megablocks to roll your own DBRX for even better performance. The new 1.6b model outperforms Gemma 2.5b & other edge models handily too, great work by @rikelhood & team. https://t.co/5UoGVefozl
New optimization method from Microsoft Research - results look pretty good.
GGUF My Repo by @huggingface Create quantum GGUF models fully online - quickly and secure. Thanks to @reach_vb, @pcuenq and team for creating this HF space! huggingface.co/spaces/ggml-or… In the video below I give it a try to create a quantum 8-bit model of Gemma 2B - it took about…
Our guy @theemozilla did work on validating and reproducing the BitNet paper with a 1B pretraining on the first 60B tokens from the @allen_ai Olmo dataset, and found very consistent results with the paper!
Our guy @theemozilla did work on validating and reproducing the BitNet paper with a 1B pretraining on the first 60B tokens from the @allen_ai Olmo dataset, and found very consistent results with the paper!
Should we acquire Stability and open-source SD3?
As my notifications are RIP some notes: 1. My shares have majority of vote @StabilityAI 2. They have full board control The concentration of power in AI is bad for us all I decided to step down to fix this at Stability & elsewhere Will be sharing more soon Exciting times
As my notifications are RIP some notes: 1. My shares have majority of vote @StabilityAI 2. They have full board control The concentration of power in AI is bad for us all I decided to step down to fix this at Stability & elsewhere Will be sharing more soon Exciting times
I was playing last few evenings with mixing a few popular #SDXL models that worked well with reference #SDXL_Lightning workflow (8 steps). I liked the results, so... released it as a MoonRide Light Mix 1 - you can check it out on #CivitAI: civitai.com/models/22511?m…. Prompts in ALT.
Looks like really good idea when multiple finetuned models are available, and we want to accumulate their knowledge and skills. I think it could work for SD models, too. Imagine mixing finetuned #SD3 variants using this method ^^.
Looks like really good idea when multiple finetuned models are available, and we want to accumulate their knowledge and skills. I think it could work for SD models, too. Imagine mixing finetuned #SD3 variants using this method ^^.
Ella-louise Koeppen @KoeppenEll77947
91 Followers 5K Following 25 - Michigan - Ella-louise💋 - TOP9.52% ofNelda Ambrosius @NAmbrosius38147
18 Followers 3K FollowingMelissa Martinez @MelissaMar29881
8 Followers 746 FollowingSamson Ibarra @IbarraSams62906
11 Followers 715 FollowingHazel @hazelhumphreys6
198 Followers 3K FollowingAnna @anna_estrada48
361 Followers 3K FollowingKristina Ulicna, PhD .. @KristinaUlicna
14K Followers 12K Following Research Scientist 👩💻 @Valence_AI 💻🧪, powered by @RecursionPharma 🔬🧬 | Visiting @Mila_Quebec in Montréal 🇨🇦 | 👩🔬 Into single cells 🧫 via #python 🐍TechAI @Tech_AI_Tech
6K Followers 7K Following Inspired by AI Data, Powered by Supposition: Unchain The Potential of Tech AI. #TechAI #LearningAI #AI #Technology #GenerativeAIHarsh Maheshwari @HarshMheshwari
1K Followers 1K Following Enthusiastic about #GenerativeAI #DataScience 🤖 | Constantly curious learner 🌱 | Applied scientist 2 at @amazon | Writer at @medium | @IITKGP GraduateRetta Pinedo @pinedo_ret31159
71 Followers 5K FollowingCaitlin Zhang @ZhangCaitl21823
9 Followers 938 FollowingRobertaBoyle @wx5WBwCg69Jw9u3
4 Followers 171 FollowingHayven ♡ @adventuwe
102 Followers 539 Following A person who likes outdoor adventures, tech, languages, 80s anime, and good books. 30+. Naturalist, anti-war, pro AI, posthumanist. (agender; they/them)ゆう@小さな起�.. @yuugaku
560 Followers 892 Following 30代のIT起業家🖥|様々なAIアートを作成しています✏️|無言フォロー歓迎。どうぞお気軽にフォロー下さい💡 I am an IT entrepreneur in my 30s🖥|I create various AI art. ✏️|Please feel free to follow me 💡Alexandra Heimann @AlexandraH19319
76 Followers 5K FollowingGrzegorz @GrzegorOryginal
74K Followers 76K Following Imagination is more important than knowledge. 🔥NO DMZainab Sirois @SiroisZain
67 Followers 5K FollowingAthena Denis @ADenis43249
63 Followers 5K FollowingJusta Hoven @JustaH95017
37 Followers 5K FollowingAnnamaria Till @AnnamariaT62870
71 Followers 5K FollowingSasha Hamiton @hamiton_sa49316
87 Followers 5K FollowingMarco Matthies @MarcoMatthies
91 Followers 2K Following Interested in math, programming, computational biology, AI, and investing.Lavonna Shansky @LavonnaSha54870
48 Followers 5K Followingpk3bcm0gcq @da9sjwhbr
3 Followers 120 FollowingHole Systems @hole_systems
267 Followers 494 Following Your living memory. Follow @heytap_tech for hardware!柊孝枝 @zhngxio33489994
0 Followers 133 FollowingAIformedicine @ai4medicine4
98 Followers 1K FollowingKeira Montesi @KeiraMonte60602
38 Followers 5K FollowingVintage Vibes @VintageVibesArt
356 Followers 864 Following Multiracial writer/artist creating AI & digital art--hand-drawn art coming soon as well. My writing account: https://t.co/ASDcoyMBXZGeeknik's {{☀️}} .. @geeknik
14K Followers 4K Following Principal Vulnerability Researcher at spiderSilk. I turn keystrokes into pixels, like code on canvas. Salsa farmer. Firefox Dev. Views & code = my own.YAMIVA @YAMIVA_
829 Followers 1K Following I upload #AIART. I use some of it in my #youtube illustrated essays. https://t.co/uJvW0gAdlB is the most recent oneEmilie Henscheid @EmiliHensch
86 Followers 5K FollowingAI Generated Producti.. @AIGenPro
635 Followers 2K Following We’re a media company that specializes in high quality entertainment generated by Artificial Intelligence. More: @GenAIChad @AIGenStudiosManie Veiga @manie_ve
49 Followers 5K FollowingTawny Recksiek @recksiek55135
66 Followers 5K FollowingDragonQuix @DragonQuix
3 Followers 132 FollowingAI Gen Artwork @AIGenArtwork
7K Followers 7K Following AI Generated Art Enthusiast #AIArt #AIArtwork #AICommunity #AIGenerated #aianimationLayan Bergman @bergm_laya
44 Followers 5K FollowingIgor Carron @IgorCarron
5K Followers 6K Following CEO https://t.co/b9fz6WvhTx @LightOnIO Paris Machine Learning Meetup (8200+) @ParisMLGroup https://t.co/jY1eeMkqJE (10M+ pageviews) @NuitBlog Rocket ScientistNylah Alimo @NylahA11509
31 Followers 5K FollowingAdelynn Cabanilla @AdelynnC7485
68 Followers 5K FollowingBrea Heinl @BreaHeinl75151
64 Followers 5K FollowingAnnie @RohitPa45519300
7 Followers 171 Following I am a Japanese. I am both eager for love and afraid of love. I have my own career. I own a cafe, restaurant, clothing store, and electronic tradingRachel Tin @RachelT35224
53 Followers 5K Followingyeswondwerr @yeswondwerr
571 Followers 297 FollowingMartin Shkreli (e/acc.. @wagieeacc
99K Followers 8K Following despite all my ragie I'm still just a wagie in a cagie working on DL Software: https://t.co/FVn3NRNrLe https://t.co/CgaoMfhUHdKanekoaTheGreat @KanekoaTheGreat
761K Followers 2K Following Banned by Vijaya • Resurrected by Elon • Independent JournalistElla Irwin @ellagirwin
39K Followers 247 Following Product Integrity leader @ Stability AI, co-host of The CryRoom Podcast, dog lover. Views expressed on X are my own.Christian Dallago @sacdallago
1K Followers 585 Following 🏳️🌈 @NVIDIAEU & @TU_Muenchen. Was @allianz, @vant_ai. BioCS+ML dude. https://t.co/2uPCxFk7dl… @[email protected]Fabian Gloeckle @FabianGloeckle
167 Followers 146 Following PhD student at @AIatMeta and @EcoledesPonts with @syhw and @Amaury_Hayat, co-supervised by @wtgowers. Machine learning for mathematics and programming.Guillaume Dalle @giomdal
2K Followers 2K Following PhD in machine learning & optimization, now postdoc at EPFL. Julia language enthusiast. Amateur songwriter (aka PianoHamster). OCD survivor.Senator Scott Wiener @Scott_Wiener
101K Followers 1K Following CA State Senator. Chair, Budget Committee. Former Chair, Legislative LGBTQ Caucus. Housing/transit/climate/criminal justice reform/health. Democrat.🏳️🌈✡️Melinda B. Chu @MelindaBChu1
4K Followers 2K Following Stealth Startup▪️VC Scout (DM for info)▪️@princeton 🐯, MD @SLU_Official 👩🏻⚕️, MBA @WUSTLBusiness 👩🏻💻Chris Lengerich /dd @chrislengerich
701 Followers 408 Following How do we make scientists and engineers 100x more productive to solve problems that matter? Systematic breakthroughs, open finance, science & psych @ContextFundBram @BramVanroy
1K Followers 712 Following @ku_leuven @ccl_kuleuven: Creative #NLG 🖋️ @ivdnt: Dutch #NLProc and #LLMs 🤖 Organizing @ctt2024 🖋️ Fellow at @huggingface 🤗 Prev. @lt3ugent, @SignONRémi 〰️ @remilouf
6K Followers 1K Following LLMs & structured generation @dottxtai. @OutlinesOSS 〰️ . Alumni @ENS_ULM & @UniOfOxford. I wander.Olivier Blanchard @ojblanchard1
170K Followers 389 Following Robert Solow Professor of economics emeritus, MIT Senior Fellow, Peterson Institute for International EconomicsErik Brynjolfsson @erikbryn
209K Followers 4K Following Director @DigEconLab Professor @StanfordHAI @SIEPR @Stanford @NBERPubs https://t.co/D2bPyxoFEfBrian Hie @BrianHie
5K Followers 403 Following Assistant professor @StanfordEng ChemE and @StanfordData, Innovation Investigator @arcinstitute | Machine learning for biologyEllen Zhong @ZhongingAlong
7K Followers 902 Following Assistant Professor @PrincetonCS. #proteins247 #ml #cryoem #compbio #cryodrgn ❄️🐉#ai4science Prev: @MIT @DeepMind @DEShawResearchZeming Lin @ebetica
2K Followers 363 Following PhD @ NYU. ESMFold / PyTorch. Climbs all day. Definitely not s̶k̶y̶n̶e̶t̶GPT5. Unsupervised learner but sometimes still gets a few rewards.Roshan Rao @proteinrosh
2K Followers 578 Following he/him. Proteins, evolutionary models, unsupervised learning. Prev: RS @MetaAI, PhD @berkeley_ai.Ali Madani @thisismadani
6K Followers 1K Following Founder & CEO of Profluent (https://t.co/GpoVdKW1zQ, we're hiring!). AI+Biology to cure disease. Berkeley PhD. formerly Research @ Salesforce AI.Kristina Ulicna, PhD .. @KristinaUlicna
14K Followers 12K Following Research Scientist 👩💻 @Valence_AI 💻🧪, powered by @RecursionPharma 🔬🧬 | Visiting @Mila_Quebec in Montréal 🇨🇦 | 👩🔬 Into single cells 🧫 via #python 🐍TechAI @Tech_AI_Tech
6K Followers 7K Following Inspired by AI Data, Powered by Supposition: Unchain The Potential of Tech AI. #TechAI #LearningAI #AI #Technology #GenerativeAIHarsh Maheshwari @HarshMheshwari
1K Followers 1K Following Enthusiastic about #GenerativeAI #DataScience 🤖 | Constantly curious learner 🌱 | Applied scientist 2 at @amazon | Writer at @medium | @IITKGP GraduateJon Askonas ⏩🚀�.. @JonAskonas
5K Followers 2K Following Politics @CatholicUniv | Senior Fellow at @joinFAI | Effective Optimist at @aftfuture | 🇺🇸/accStephanie Chan @scychan_brains
3K Followers 2K Following Senior Research Scientist at DeepMind. Artificial and biological brains 🤖 🧠 Views are my ownTed Moskovitz @ted_moskovitz
743 Followers 193 Following PhD student at @GatsbyUCL. Formerly: intern at @DeepMind, @UberAILabs, student at @ColumbiaCompSci, @PrincetonNeuro.Ahmed Khaled @akhaledv2
501 Followers 771 Following I like optimization, mathematics, and drawing. Ph.D. student at Princeton ECE. Pronouns: he/him.tsvetshop @tsvetshop
800 Followers 131 Following Group account for Prof. Yulia Tsvetkov's lab at @uwnlp. We work on low-resource, multilingual, social-oriented NLP. Details on our website:Vidhisha Balachandran @vidhisha_b
521 Followers 490 Following Researcher @MSFTResearch, PhD from @LTIatCMU, Ex-Intern @allen_ai, @GoogleAI | NLP/AI | she/herToviah Moldwin @TMoldwin
673 Followers 832 Following CEO and founder, https://t.co/AkHPAGcWFQ. Computational neuroscientist @ELSCbrain @Segev_Lab. Singer and guitarist for the rock band @SynfireChain. Dualist.Richard Futrell @rljfutrell
2K Followers 728 Following Language Science at University of California, Irvine Information theory and languageKabir @kabirahuja004
452 Followers 419 Following CSE PhD Student @uwnlp | Ex-RF @MSFTResearch | cinephile 🎥Aaron Mueller @amuuueller
749 Followers 548 Following Postdoc with @boknilev and @davidbau ≡ PhD from @jhuCLSP ≡ Into #NLProc, interpretability, and computational psycholinguistics 💻🧠David Bau @davidbau
3K Followers 242 Following Computer Science Professor at Northeastern, Ex-Googler. Believes AI should be transparent. @[email protected] @davidbau.bsky.social https://t.co/wmP5LUZRTwSimone Balloccu @simoneballoccu
272 Followers 226 Following (he/him) Postdoc @ufal_cuni, ERC NG-NLG project / interested in AI applied to behaviour change; NLP for mental health; NLG in general.Jonathan Herzig @jonherzig
318 Followers 371 FollowingAdi Simhi @AdiSimhi
83 Followers 93 FollowingMax Roser @MaxCRoser
287K Followers 1K Following Data to understand global problems and research to make progress against them. Founder of @OurWorldInData / Professor at @UniofOxford's @BlavatnikSchoolEric J. Michaud @ericjmichaud_
1K Followers 775 Following PhD student at MIT. Trying to make deep neural networks among the best understood objects in the universe. 💻🤖🧠👽🔭🚀Samuel Marks @saprmarks
713 Followers 79 Following Postdoc studying interpretability for AI safety under @davidbau. PhD in math from @harvard. Previously director of technical programs at https://t.co/FxRv4QgERO.Mostafa Elhoushi @m_elhoushi
584 Followers 1K Following Research Engineer at Work. Volunteering for Various Causes after Work. Opinions are my own. 🇵🇸Aaditya Singh @Aaditya6284
447 Followers 247 Following PhD student at @GatsbyUCL working with @SaxeLab, @FelixHill84 on learning dynamics, ICL, concepts, LLMs. Prev. at: @GoogleDeepMind, @AIatMeta (LLaMa 3), @MITSean Blagsvedt @SeanB
1K Followers 503 Following Founder https://t.co/IKovPp0ank • Past: https://t.co/6rPLA053T2, https://t.co/4Jl2CltKX6, Marco Polo, Microsoft Research India, White House • Dad, Feminist, CoderPratik Desai @chheplo
9K Followers 708 Following 🌾 KissanAI - Pioneering Vernacular AgriCoPilot Platform with Agri Vertical Model (Dhenu) for Climate Resilient Agriculture (Kissan=Farmer)Baptiste Rozière @b_roziere
1K Followers 181 Following Research scientist at Meta AI. Working on ML for programming languages.LLM4Code @llm4code
331 Followers 101 Following The 1st International Workshop on Large Language Models for Code Co-located with @ICSEconf 2024The first randomized trial of medical #AI to show it saves lives ECG-AI alert in 16,000 hospitalized patients 31% reduction of mortality (absolute 7 per 100 patients) in pre-specified high-risk group nature.com/articles/s4159… @NatureMedicine
Open letter to @Scott_Wiener re: SB-1047. A Safe Harbor for Independent AI Evaluation? Hi Scott, Just a personal thought from the investing perspective, 1047 seems likely to be about 1 week - 6 months away from an SBF 2.0-style scandal. The bill sponsors likely aren't being…
I thought it was a bad idea to seize Russian reserves before the US congress had voted on the Ukraine package. It gave too easy a way to Congress to vote no and pass the buck. Now that they have voted, it is hard to think of good reasons not to seize. Yes, it will create a…
We are already seeing an explosion of AI regulation that is designed to ban open source while claiming to be neutral. SB 1047 designates a "hazardous capability" to include what a third party can show with infinite fine-tuning and re-training. Meanwhile, closed models get points…
The thread also contains fear mongering about open source. Unless you think catastrophic harms are going to happen left & right with open source projects, it’s absolutely false to say this “de facto criminalizes open source.” + these are civil sanctions, not criminal liability!
The new california law gets dumber the more you read. if anyone can continue training and finetuning your model to make it cause harm, then it's considered to have a hazardous capability. with this law in place it's arguable that you couldn't even release untrained weights.
We are already seeing an explosion of AI regulation that is designed to ban open source while claiming to be neutral. SB 1047 designates a "hazardous capability" to include what a third party can show with infinite fine-tuning and re-training. Meanwhile, closed models get points…
We've been in the kitchen cooking 🔥 Excited to release the first @AIatMeta LLama-3 8B with a context length of over 1M on @huggingface - coming off of the 160K context length model we released on Friday! A huge thank you to @CrusoeEnergy for sponsoring the compute. Let us know…
Almost no representation from startups, or open source except @drfeifei. Do us proud Fei Fei. You're our only hope!
This morning the Department of Homeland Security announced the establishment of the Artificial Intelligence Safety and Security Board. The 22 inaugural members include Sam Altman, Dario Amodei, Jensen Huang, Satya Nadella, Sundar Pichai and many others.
📢 New Paper! Ever wondered why transformers are able to capture hierarchical structure of human language without incorporating an explicit 🌲 structure in their architecture? In this work we delve deep into understanding hierarchical generalization in transformers. (1/n)
Can you imagine having all the evidence of data contamination gathered in one place? 📢As part of the CONDA workshop, we present the Data Contamination Evidence Collection, a shared task on reporting contamination. Available as a @huggingface space: hf.co/spaces/CONDA-W…
The Data Contamination Database continues to grow!!! We aim to build a centralized collection of data contamination evidence, allowing the community to easily verify if a model/dataset has been contaminated. You can contribute by submitting a PR! huggingface.co/spaces/CONDA-W…
Our new preprint, "Constructing Benchmarks and Interventions for Combating Hallucinations in LLMs", provides some guidance. This work was done with @jonherzig, Idan Szpektor and @boknilev Check out our work here: arxiv.org/abs/2404.09971
Myth: open foundation models are antithetical to AI safety. Fact: open foundation models are critical for AI safety. Here are three reasons why:
Releasing StarCoder2 Instruct! 🚀 Achieves 72% HumanEval score using only self-generated content without any GPT-3.5/4 data. This work demonstrates that self-instruct works already well at the 15B scale without data from proprietary models! Read more: huggingface.co/blog/sc2-instr…
Feeling proud to not be a california resident today lol
🔊By plugging ur own motion masks and adjusting Controlnet values you can define how prominent the motion mask should be and the overall flow. This motion mask I made in #Touchdesigner to get a beatsynced animation. @1null1 (Free for non-commercial). No post work, from Comfy:
California Bill 1047 has been fasttracked: • Covers all models made w/ 10^26 flops • Covers all models with similar perf to above • Creates a Frontier Model Division to report to • Devs must assert such models are safe under penalty of perjury text: legiscan.com/CA/text/SB1047…
@NickADobos Read the peer review critiques at the bottom