Avi Caciularu @clu_avi
Research Scientist @GoogleAI | ex ML & NLP PhD student @biunlp, intern at @allen_ai, @Microsoft, @AIatMeta. aviclu.github.io Joined July 2009-
Tweets173
-
Followers371
-
Following397
-
Likes859
More exciting news today -- Gemini 1.5 Pro result is out! Gemini 1.5 Pro API-0409-preview now achieves #2 on the leaderboard, surpassing #3 GPT4-0125-preview to almost top-1! Gemini shows even stronger performance on longer prompts, in which it ranks joint #1 with the latest…
More exciting news today -- Gemini 1.5 Pro result is out! Gemini 1.5 Pro API-0409-preview now achieves #2 on the leaderboard, surpassing #3 GPT4-0125-preview to almost top-1! Gemini shows even stronger performance on longer prompts, in which it ranks joint #1 with the latest… https://t.co/IvrwwMRhkU
Thrilled to share our latest blogpost with @ghandeharioun on 🩺Patchscopes, our groundbreaking framework that decodes hidden information from LLMs. Dive into our blog post to explore how Patchscopes is setting a new standard in LLM interpretability. #NLProc #AI #MachineLearning
Thrilled to share our latest blogpost with @ghandeharioun on 🩺Patchscopes, our groundbreaking framework that decodes hidden information from LLMs. Dive into our blog post to explore how Patchscopes is setting a new standard in LLM interpretability. #NLProc #AI #MachineLearning
🐣Too Long; Didn't Read: Most LLMs cite full documents for support, but they're too long so people don't actually read them. In our new work "Attribute First, then Generate”, we introduce fine-grained attributions, where LLMs are required to highlight only relevant information…
Need a break from ICML/ACL rebuttals? This activity is for you Many people work on model analysis and interpretability research these days, but is such research helpful for deriving progress in NLP? Help us answer this question by filling out our survey: forms.gle/XVbcLobx7j1kEZ…
We recently found that few-shot tool-usage algorithms, like RAG and invoking calculators, often perform not better than letting the LM operate w/o tools. We wrote a blogpost summarizing key findings and takeaways for future work on tool-use, check it out! medium.com/@alonjacovi/fe…
Delving into the enigmatic world of text compression & tokenization in the era of LLMs has been a fun journey, all thanks to @omerNLP 's ideas 🚀. Grateful for the opportunity to explore these complexities and share his invaluable insights 🤓
Delving into the enigmatic world of text compression & tokenization in the era of LLMs has been a fun journey, all thanks to @omerNLP 's ideas 🚀. Grateful for the opportunity to explore these complexities and share his invaluable insights 🤓
What’s the plan: Our study demonstrates that LLMs lack necessary skills required for planning, in-line with previous research showing that they are unable to generate executable plans. We also suggest techniques on how to improve! 👇 (1/6)
👋 Check out our new paper and benchmark: REVEAL, a dataset with step-by-step correctness labels for chain-of-thought reasoning in open-domain QA 🧵🧵🧵 arxiv.org/abs/2402.00559 huggingface.co/datasets/googl…
We're organizing a new workshop on data contamination (CONDA) at ACL2024! CFP below Contamination is when test data was included in training, for any reason. A lot of research is needed to understand when this happens, or prevent it. Consider submitting! x.com/osainz59/statu…
We're organizing a new workshop on data contamination (CONDA) at ACL2024! CFP below Contamination is when test data was included in training, for any reason. A lot of research is needed to understand when this happens, or prevent it. Consider submitting! x.com/osainz59/statu…
This is a really wonderful paper that I have yet to read through fully, but I think anyone into linguistics should pay close attention to table 3, where they show how the transformer hierarchically builds up a representation of a noun phrase layer-by-layer!
This is a really wonderful paper that I have yet to read through fully, but I think anyone into linguistics should pay close attention to table 3, where they show how the transformer hierarchically builds up a representation of a noun phrase layer-by-layer! https://t.co/5oaFpm3Ncd
Thrilled to share our collaborative effort on 🩺Patchscopes🩺, a new general framework for better interpreting LLMs via decoding specific information from representations 🚀. Dive deeper into our findings and results in @ghandeharioun's insightful 🧵 #LLMInterpretability#NLProc
Thrilled to share our collaborative effort on 🩺Patchscopes🩺, a new general framework for better interpreting LLMs via decoding specific information from representations 🚀. Dive deeper into our findings and results in @ghandeharioun's insightful 🧵 #LLMInterpretability#NLProc
Google presents Patchscopes A Unifying Framework for Inspecting Hidden Representations of Language Models paper page: huggingface.co/papers/2401.06… Inspecting the information encoded in hidden representations of large language models (LLMs) can explain models' behavior and verify…
New preprint!💡 We examine how multilingual instructions and responses affect open-ended instruction-following across languages.🌎 @jonherzig @roeeaharoni Idan Szpektor @rtsarfaty @mataneyal1 @GoogleAI 📜arxiv.org/abs/2401.01854 👇 1/5
New preprint!💡 We examine how multilingual instructions and responses affect open-ended instruction-following across languages.🌎 @jonherzig @roeeaharoni Idan Szpektor @rtsarfaty @mataneyal1 @GoogleAI 📜arxiv.org/abs/2401.01854 👇 1/5
Research Opportunities @Google-Please RT Excited to be at #NeurIPS2023! If interested in how model #interpretability can(not) lead to better control, improving our #mechanistic understanding of Transformers & thinking about what the future models could look like, let’s chat!(1/n)
Ever wondered if LLMs understand when they don’t have enough information to answer a question even if they hallucinate something random? Come check out our poster that addresses just that (East Foyer @emnlpmeeting ) w\ @ravfogel @omerNLP @clu_avi
Introducing Gemini 1.0, our most capable and general AI model yet. Built natively to be multimodal, it’s the first step in our Gemini-era of models. Gemini is optimized in three sizes - Ultra, Pro, and Nano Gemini Ultra’s performance exceeds current state-of-the-art results on…
On my way to #EMNLP2023 to present 3 of my papers - (1 demo, 1 findings and 1 main conference - see thread👇). Let’s catch up if you are there!
probing is fun, right? but also a bit of a mess, right? Join me for a 🚨new paper🚨 thread! "Is Probing All You Need? Indicator Tasks as an Alternative to Probing Embedding Spaces" with Tal Levy and @rtsarfaty to appear in #EMNLP findings and in the @BlackboxNLP workshop. 1/🧵
Worth checking out 🚀
Leshem Choshen 🤖�.. @LChoshen
4K Followers 548 Following 🥇 Collaborative LLMs 🥈 Opinionatedly sharing #ML & #NLP 🥉 Propagating us underdogs we owe science an alternative hype @IBMResearch & @MIT_CSAILOr Honovich @OHonovich
197 Followers 223 FollowingOri Ram @ori__ram
765 Followers 386 Following Research Scientist @GoogleAI, working on #NLProc. Previously: PhD from @TelAvivUni, Research Scientist @AI21LabsOfir Press 🖋 @OfirPress
10K Followers 3K Following I build tough benchmarks for LMs and then I get the LMs to solve them. Postdoc @Princeton. PhD from @nlpnoah @UW. Ex-visiting researcher @MetaAI & @MosaicML.Natalie Shapira @NatalieShapira
1K Followers 225 Following Tell me about challenges, the unbelievable, the human mind and artificial intelligence, thoughts, social life, family life, science and philosophy.Uri Shaham @Uri_Shaham
365 Followers 242 Following PhD candidate at Tel-Aviv University. Research intern at Google Research.roeeaharoni @roeeaharoni
2K Followers 758 Following Research Scientist @GoogleAI Tel-Aviv | Phd @BIUNLP LabYonatan Bitton @YonatanBitton
923 Followers 924 Following Research Scientist at @Google CS PhD at @HebrewU. Recent research areas include image-text alignment, text-to-Image models, and visual instruction tuning.Ben Bogin @ben_bogin
630 Followers 421 Following CS PhD student at Tel-Aviv University, studying #NLProc. https://t.co/LPRm6GDjvtArie Cattan @ArieCattan
352 Followers 537 Following CS Phd student at @biunlp and intern at @Google, previously @IBMResearch @allen_aiMatan Eyal @mataneyal1
191 Followers 368 FollowingMaor Ivgi @maorivg
361 Followers 157 Following NLP researcher / Ph.D. candidate at Tel-Aviv UniversityValentina Pyatkin @valentina__py
2K Followers 1K Following Postdoc at the Allen Institute for AI @allen_ai and @uwnlpOren Sultan @oren_sultan
681 Followers 597 Following AI Researcher & Data Scientist @Lightricks, CS PhD Candidate #AI #NLP @HebrewU, advised by @HyadataLab 🇮🇱 | prev. @TU_Muenchen 🇩🇪 @UniMelb 🇦🇺 8200 UnitBIU NLP @biunlp
694 Followers 101 Following The Bar-Ilan University, Natural Language Processing group.Guy Dar @guy_dar1
361 Followers 221 Following #NLProc Researcher | #AI #NLProc #interpretability | opinions my own sadly | off-topic tweets erased periodically | he/himMarilyn Wetzel @MarilynWet60218
86 Followers 5K FollowingEunsol Choi @eunsolc
5K Followers 815 Following on natural language processing / machine learning. assistant professor @UTCompSci. prev @googleai, @uwcse, @Cornell. opinions are of my own.Denis Bykov @denis_bykov
144 Followers 606 Following Search @ Yango Maps. Leading talented minds in crafting innovative search engines.Mahed Mousavi | ما�.. @mahedmousavi
126 Followers 178 Following Junior Assistant Professor & Research Fellow (RTD-A) Computational Linguistics University of Trento (cover photo @herbertgreg8)Joe Stacey @_joestacey_
569 Followers 1K Following PhD student at Imperial and Apple Scholar. I love running, NLP and travelling (in no particular order). Ex teacher and PwC Consultant. #NLProcPensé FFun @inftyCategory
99 Followers 6K FollowingItxaso Baskero Dorrea.. @IDorreak
12 Followers 365 FollowingOlia Toporkov @oaizkora
12 Followers 26 Following PhD student in NLP, HiTZ-Center-Ixa, University of the Basque CountryArif Ahmad @arif_ahmad_py
277 Followers 7K Following All things AI, Computer Science and Circuits! Prev. @GoogleAIGuy Kaplan @GKaplan38844
3 Followers 88 FollowingArjun Srivastava @arjunsriv
63 Followers 1K Following AI, reinforcement learning, distributed systems something new @Woven_ToyotaJP prev - discovery @bookmyshow, cs @IITIOfficialZhaoyi LI @LiZhaoyi13
5 Followers 120 Following CS PhD student at University of Science and Technology of China @ustcglobal,@studyatustc & CityU of Hong Kong @cityuhongkong. NLP research.Satvik Dhandhania @satvikd22
104 Followers 961 Following @Hubspot , CX advisor@UCI, @99tartans, @meta, @Microsoft, @motorolaus, @carnegieMellon @VIT_univ, #tech #engineering #startups #investingAlexander Wan @alexwan55
475 Followers 944 Following CS at Berkeley; @BerkeleyML @BerkeleyNLP; NLP researchAshutosh Mehra @ashutoshmehra
1K Followers 5K Following Senior Principal Scientist at Adobe. Working on Acrobat AI Assistant, LLMs, and document ML.Yolando Loiko @yoland_loi
43 Followers 5K FollowingAviv Gelfand @BGlpnd
18 Followers 92 FollowingOfec Israel @ofecisrael
13 Followers 152 FollowingAI-llama @gaoqiang_nlp
5 Followers 70 Following a postgraduate student with focus on nlp from WhuHan university . i will graduate at 2025 and i’m looking for a phd position.Tom @Tom59159284
39 Followers 396 FollowingShaltiel @SShmidman
161 Followers 60 FollowingYoung @younqchan
170 Followers 3K Following Final year Ph.D. student working on Out-of-Distribution Generalization and Causality of Large Pre-trained Models, and Graph Neural Networks.Ben Hagag @BenHagag20
44 Followers 110 Following We see only what we know 🙊🙈🙉 | Head of Research @DarrowAI & Researcher at #BIU | Exploring the intersections of #NLProc, #LegalNLP, and #JustNLPMichael Toker @michael_toker
37 Followers 362 Following PhD candidate @Technion NLP lab - Developing explainability methods to gain a better understanding of LLMsSanchit Ahuja @SanchitAhuja7
369 Followers 862 Following Trynna work. Research Fellow at @MSFTResearch x-ml at @SkitTech Alum at @bitspilaniindia.Kris Cao @kroscoo
1K Followers 647 Following When lava pours out near the sea's surface tremendous volcanic explosions sometimes occur | Research scientist @DeepMind working on languageJuan Hmmm @JuanAH03488233
79 Followers 3K FollowingJuliana Strahm @JulianaStr31778
28 Followers 2K Following 🔑Juliana | 20 | Earn your own Crypto casino👇⚡Sasha Goldshtein @goldshtn
4K Followers 1K Following Software Engineer at Google Research. I work on Gemini factuality. Opinions my own. He/him.Amirhossein Abaskohi @AmirAbaskohi
145 Followers 871 Following Master Student @UBC_CS | NLP Researcher @UBC_NLP | Content Creator @YouTube and @Medium #NLProc #MachineLearningRomano @____romano____
182 Followers 664 Following Praise the Lord and pass the ammunition. God wants you to go to war.Shir @ShirAshuryTahan
20 Followers 106 Following MSc Student at @biunlp and NLP Reseacher at @ibmresearchSunset @Sunset16089641
1K Followers 4K Following Greetings. Lead Programmer of S.E.T. Group. PG-13 (doing my best). Always not online. Backup account: @sunset_backupTrung Tran Thanh (T) @TrungTT10
40 Followers 326 Following co- founder of https://t.co/IdqlIIs4uK | Computer Vision specialist | NLP researcher | CTO of ClientScanYi Lin Sung @yilin_sung
521 Followers 730 Following CS PhD student @unccs @uncnlp | Previously intern @MetaAI @MSFTResearch | Multi-modal DL, Efficient fine-tuning.Emily Reynolds @EmilyReyno85185
70 Followers 3K FollowingAl Mamun @al_mamun_sardar
276 Followers 3K Following Looking for NLP/LLM related PhD positions | Research Assistant (NLP) at Jahangirnagar University | MSc (CS)Rushikesh Zawar @ZawarRushikesh
15 Followers 263 Followingpawann k. @pawaniiit
221 Followers 4K Following Prof., PhD, Inria, France, Postdoc KU Leuven, Fraunhofer ITWM, FU Berlin. I like Machine learning and mathematics.Eyal Orbach @eyalOrbach
62 Followers 261 FollowingMyron @myrondza10
40 Followers 2K Following Data Scientist 🧠😍✨ Game-changing AI is here! | ML / AI + Neuroscience + Robotics(((ل()(ل() 'yoav))).. @yoavgo
46K Followers 2K FollowingLeshem Choshen 🤖�.. @LChoshen
4K Followers 548 Following 🥇 Collaborative LLMs 🥈 Opinionatedly sharing #ML & #NLP 🥉 Propagating us underdogs we owe science an alternative hype @IBMResearch & @MIT_CSAILAK @_akhaliq
310K Followers 3K Following AI research paper tweets, ML @Gradio (acq. by @HuggingFace 🤗) dm for promo follow on Hugging Face: https://t.co/q2Qoey80GxOr Honovich @OHonovich
197 Followers 223 FollowingOri Ram @ori__ram
765 Followers 386 Following Research Scientist @GoogleAI, working on #NLProc. Previously: PhD from @TelAvivUni, Research Scientist @AI21LabsOfir Press 🖋 @OfirPress
10K Followers 3K Following I build tough benchmarks for LMs and then I get the LMs to solve them. Postdoc @Princeton. PhD from @nlpnoah @UW. Ex-visiting researcher @MetaAI & @MosaicML.Natalie Shapira @NatalieShapira
1K Followers 225 Following Tell me about challenges, the unbelievable, the human mind and artificial intelligence, thoughts, social life, family life, science and philosophy.Uri Shaham @Uri_Shaham
365 Followers 242 Following PhD candidate at Tel-Aviv University. Research intern at Google Research.roeeaharoni @roeeaharoni
2K Followers 758 Following Research Scientist @GoogleAI Tel-Aviv | Phd @BIUNLP LabYonatan Bitton @YonatanBitton
923 Followers 924 Following Research Scientist at @Google CS PhD at @HebrewU. Recent research areas include image-text alignment, text-to-Image models, and visual instruction tuning.Ben Bogin @ben_bogin
630 Followers 421 Following CS PhD student at Tel-Aviv University, studying #NLProc. https://t.co/LPRm6GDjvtArie Cattan @ArieCattan
352 Followers 537 Following CS Phd student at @biunlp and intern at @Google, previously @IBMResearch @allen_aiMatan Eyal @mataneyal1
191 Followers 368 FollowingMaor Ivgi @maorivg
361 Followers 157 Following NLP researcher / Ph.D. candidate at Tel-Aviv UniversityAmir Feder @amir_feder
509 Followers 189 Following Causal inference and ML with text. Postdoc @blei_lab // @DataSciColumbia | @GoogleAISivan Doveh @SivanDoveh
179 Followers 299 Following CS Ph.D. Candidate @ Weizmann & IBM Research; Mainly Vision&LanguageKris Cao @kroscoo
1K Followers 647 Following When lava pours out near the sea's surface tremendous volcanic explosions sometimes occur | Research scientist @DeepMind working on languageMosh Levy @mosh_levy
267 Followers 163 Following phd student @biunlp. studying ai robustness and behaviors.Sasha Goldshtein @goldshtn
4K Followers 1K Following Software Engineer at Google Research. I work on Gemini factuality. Opinions my own. He/him.Vinh Q. Tran @vqctran
1K Followers 282 Following i research language models @Google, all thoughts my own, he/himRock Mada @RackMada
31K Followers 1K Following נשוי, אבא, חוקר, דוקטור, מנסה לספר על מדע ישן וחדש ! Scientist , share science ״מכורה שלי, ארץ נוי אביונה״Computer Science @CompSciFact
249K Followers 19 Following Daily tweets about computer science and related stuff @JohnDCook.The PhD Place @ThePhDPlace
107K Followers 18K Following Empowering doctoral researchers through community 🎓 Weekly Guides @DoctoralStories 📚 Online writing group @AcWriClub ✍️ #AcademicTwitterPeter Hase @peterbhase
2K Followers 691 Following Google PhD Fellow at @uncnlp. Interested in interpretable ML, natural language processing, AI Safety, and Effective Altruism.Jasmijn Bastings @jasmijnbastings
4K Followers 2K Following Sr Research Scientist @GoogleDeepMind. Interested in gender, feminism, fairness, bias & ethics in #NLProc/#AI. Views my own. She/they.Adam Pearce @adamrpearce
6K Followers 372 Following @anthropicai, previously: google brain, @nytgraphics and @bbgvisualdataAhmad Beirami @abeirami
4K Followers 2K Following Building safe, helpful, and scalable generative AI @Google | ex-{@AIatMeta, @EA, @MIT, @Harvard, @DukeU} | @GeorgiaTech PhD | زن زندگی آزادی | opinions my ownAkari Asai @AkariAsai
11K Followers 650 Following Ph.D. student @uwcse & @uwnlp. NLP. IBM Ph.D. fellow (2022-2023). Meta student researcher (2023-) . ☕️ 🐕 🏃♀️🧗♀️🍳Percy Liang @percyliang
49K Followers 408 Following Associate Professor in computer science @Stanford @StanfordHAI @StanfordCRFM @StanfordAILab @stanfordnlp | cofounder @togethercompute | PianistSara Hooker @sarahookr
39K Followers 7K Following I lead @CohereForAI. Formerly Research @Google Brain @GoogleDeepmind. ML Efficiency at scale, LLMs, @trustworthy_ml. Changing spaces where breakthroughs happen.Asma Ghandeharioun @ghandeharioun
2K Followers 489 Following Research Scientist @GoogleAI working on ML interpretability & human-centered AI, PhD from @MITGal Yona @_galyo
418 Followers 432 Following Research scientist @googleai, previously CS PhD @weizmannscienceHadas Orgad @OrgadHadas
165 Followers 90 Following PhD student (Natural Language Processing) @ Technion, Israel, Interested in AI interpretability, robustness and safetyBrian Gordon @Brian_Gordon13
27 Followers 63 FollowingEyal Ben-David @bd_eyal
90 Followers 147 Following Research scientist @ Google Tel Aviv | PhD @ TechnionFederico Baldassarre @BaldassarreFe
104 Followers 402 Following Postdoc @AIatMeta, PhD @kth_rpl. Deep learning explainability, concept-based representations, and reasoning in computer vision.Kara Swisher @karaswisher
1.5M Followers 2K Following “Vitriolic” and now “shrill”media lady, though dogs can hear me loud and clearroi kais • روعي.. @kaisos1987
82K Followers 5K Following ראש תחום העולם הערבי של כאן 11 | Arab Affairs Correspondent Kan | هيئة الإذاعة والتلفزيون الإسرائيليّة كان 11 [email protected]Amir Hetsroni @AmirHetsroni
16K Followers 279 Following Professor of communication, free speech and pro-vaccination activist.Kelvin Guu @kelvin_guu
3K Followers 333 Following Senior staff research scientist @ Google DeepMind leading cross-functional teams of 40+ (research/eng/PM/UI/UX), turning our SOTA research into new AI products.Nitzan Barzilay @Nitzan_Barzilay
93 Followers 161 Following CSE MSc #NLProc at @nlphuji. Goes everywhere with Greg, future service dog for the blind 🦮Linoy Tsaban🎗️ @linoy_tsaban
2K Followers 893 Following Exploring the world of AI Art as a ML engineer @HuggingFace 🤗 | ✡️ & 🇮🇱 #BringThemHome 🎗️Stephanie Chan @scychan_brains
3K Followers 2K Following Senior Research Scientist at DeepMind. Artificial and biological brains 🤖 🧠 Views are my ownStability AI @StabilityAI
190K Followers 31 Following We are building the foundation to activate humanity's potential.Rishu Kumar @rishdotuk
548 Followers 538 Following A student of language @EM_LCT (@ufal_cuni & @LstSaar). Machine Translation and Summarisation #NLProcAbhijnan Nath @AbhijnanN
80 Followers 273 Following Keen observer of l'affaire politics, culture and life. Loves Chopin, Pavarotti and benign sarcasm.Felix Hill @FelixHill84
9K Followers 777 Following Research Scientist, Deepmind I try to think hard about everything I tweet, esp on 90s football and 80s music None of my opinions are really someone else'sReka @RekaAILabs
11K Followers 13 Following An AI research and product company 🫠. We are a team of scientists and engineers building state-of-the-art multimodal language models 😻Itamar Golan 🤓 @ItakGol
16K Followers 485 Following CEO & Co-founder @prompt_security ||| AI Researcher ||| LLM hackerShir Iluz @IluzShir
78 Followers 136 Following EE MSc student @ Tel Aviv University | Computer Vision & Deep Learning | Generative AI | Word-As-Image 💻👩🎨Hila Noga هيله ن.. @hila_noga
854 Followers 271 Following NLP | Tech Lead Manager @ Google Research | Personal opinions | She/her | Trust me, I’m an engineer.Samuel AMOUYAL @AmouyalSamuel
64 Followers 78 FollowingAmir David Nissan coh.. @AmirDNC
54 Followers 24 FollowingACL Mentorship @aclmentorship
2K Followers 34 Following ACL Year-Round Mentorship Program: https://t.co/F3rgnwIKUGBlackboxNLP @BlackboxNLP
371 Followers 13 Following The largest workshop on analysing and interpreting neural networks for NLP. BlackboxNLP will be held at EMNLP 2024 in Miami! Account run by @JumeletJMichael Hassid @MichaelHassid
159 Followers 88 Following PhD candidate @HebrewU; Research Assistant @MetaAI (FAIR)Leonie Weissweiler @LAWeissweiler
790 Followers 314 Following Visiting Researcher with @adelegoldberg1 at @Princeton | prev. @cislmu @LTIatCMU @CambridgeLTLZorik Gekhman @zorikgekhman
97 Followers 211 Following #NLProc PhD student @TechnionLive | Research intern @GoogleAIPete Skomoroch @peteskomoroch
53K Followers 7K Following Investor and AI startup founder. Focus: AI, LLMs, LifeOps, AI Product Management. Was founder @SkipFlag. EIR @Accel. Data Science & ML @LinkedIn, @AOL & @MITRehan Ahmed @_rehan_a
242 Followers 289 Following PhD Candidate, NLP research @BoulderNLP, Gamer, Reader, Cruciverbalist, Biryani cooker. ex-intern @explosion_aiNew paper from our team, led by @pat_verga Are you: * Doing evaluation with LLMs? * Using a huge model? * Worried about self-recognition? Try an ensemble of smaller LLMs. Use a PoLL: less biased, faster, 7x cheaper. Works great on QA & Arena-hard evals arxiv.org/abs/2404.18796
While refreshing my understanding of tokenization, I stumbled upon "Unpacking Tokenization: Evaluating Text Compression and its Correlation with Model Performance" arxiv.org/abs/2403.06265 by @omerNLP and co. Check this out!
I don't think it's productive or effective for a PhD student to ever lead more than 1 project simultaneously. If anything, I think leading 0.5 projects is even better (see SWE-bench & SWE-agent which Carlos and John co-led) Focusing is really important.
Out of curiosity, do AI PhDs normally work (lead) on several projects simultaneously? I have never managed to work on more than one project during my PhD and I tried to convince my students not to do so. The paradigm might have already changed, so I am asking here.
Lazy twitter: A common question in NLP class is "if xBERT worked well, why didn't people make it bigger?" but I realize I just don't know the answer. I assume people tried but that a lot of that is unpublished. Is the theory that denoising gets too easy for big models?
Yay 🦾 This model is really quite good, especially in instruction following: I felt a real difference when playing with it during the last few weeks.
More exciting news today -- Gemini 1.5 Pro result is out! Gemini 1.5 Pro API-0409-preview now achieves #2 on the leaderboard, surpassing #3 GPT4-0125-preview to almost top-1! Gemini shows even stronger performance on longer prompts, in which it ranks joint #1 with the latest…
@clu_avi @ghandeharioun I read the paper several months ago. I would say Patchscopes is indeed impressive and full of imagination. 😲 Its idea can help us (to some extent) alleviate the inherent shortcomings of Logit Lens when interpreting the internal hidden states of LLMs.
Being able to interpret an #ML model’s hidden representations is key to understanding its behavior. Today we introduce Patchscopes, an approach that trains #LLMs to provide natural language explanations of their own hidden representations. Learn more → goo.gle/4aS5epd
@Yampeleg Until they show some language modeling experiments you should calm down.
⌘R and, now, ⌘R+ are available now, including their weights!🥁🥁🥁 They are great combining RAG, Citing and tool-use - a really powerful combo. I am particularly passionate about grounded generation with citations/attribution (may tweet more on this soon!). Check them out!
Today, we’re introducing Command R+: a state-of-the-art RAG-optimized LLM designed to tackle enterprise-grade workloads and speak the languages of global business. Our R-series model family is now available on Microsoft Azure, and coming soon to additional cloud providers.
TL;DR: we develop methods for adding concise & accurate attributions to LLM generated texts. Each generated sentence is paired with highlighted spans from supporting sources so you can easily verify it. Check Aviv's thread and our paper for more details: arxiv.org/abs/2403.17104
🐣Too Long; Didn't Read: Most LLMs cite full documents for support, but they're too long so people don't actually read them. In our new work "Attribute First, then Generate”, we introduce fine-grained attributions, where LLMs are required to highlight only relevant information…
🐣Too Long; Didn't Read: Most LLMs cite full documents for support, but they're too long so people don't actually read them. In our new work "Attribute First, then Generate”, we introduce fine-grained attributions, where LLMs are required to highlight only relevant information…
אפשר להבין איך זה ש136 אזרחי ישראל כלואים במנהרות בעזה כבר 175 ימים, וכל מה שאנחנו מדברים עליו הוא גיוס חרדים משל אנחנו עדיין ב6 באוקטובר. אני לא יודע את מי זה משרת, אבל ברור לכם שזה ספין כן?
It was great to be back in Edinburgh! ☔️ Thank you for having me @tomsherborne @PMinervini @EdinburghNLP
Today! @valentina__py is at @EdinburghNLP @InfAtEd talking about Underspecified Language In Context
Need a break from ICML/ACL rebuttals? This activity is for you Many people work on model analysis and interpretability research these days, but is such research helpful for deriving progress in NLP? Help us answer this question by filling out our survey: forms.gle/XVbcLobx7j1kEZ…
We recently found that few-shot tool-usage algorithms, like RAG and invoking calculators, often perform not better than letting the LM operate w/o tools. We wrote a blogpost summarizing key findings and takeaways for future work on tool-use, check it out! medium.com/@alonjacovi/fe…
@clu_avi @NatalieShapira @AmirDNC הוא לקח את ג'מה 7B ועשה עליו merge עם עצמו כדי להגדיל את המודל, ולכן מודל הבסיס קטן יותר - אבל זה אותו אחד.
@clu_avi @SShmidman @AmirDNC אני מכירה לפחות עוד מודל אחד שאומן על עברית + מולטילינגוול. בכל אופן, מעניין לראות את הפער מהמודל הכי חזק באמת...
תודה לים! גם כאן אני ו@AmirDNC הרצנו בדיקות על המודל על דאטסטים קיימים בעברית, ועדיין התוצאות קצת מוזרות: פרטים טכניים: V1 הורץ עם: 1. גירסת טרנספורמרז 4.38.1 2. דיוק מלא (בלי קוונטיזציה) 3. בלי BOS V2 הורץ עם: 1. גירסת טרנספורמרז 4.38.2 2. דיוק מלא (בלי קוונטיזציה) 3. עם BOS
מודל שפה גדול, פתוח ובעברית! --- אני משחרר לכם את מודל השפה הפתוח הגדול ביותר והחזק ביותר שאי פעם אומן בעברית! (והוא מצוין גם באנגלית) --- המודל - מודל בסיס גרסה א': huggingface.co/yam-peleg/Hebr… - מודל בסיס גרסה ב' (מאומן עוד יותר): huggingface.co/yam-peleg/Hebr… - מודל המאומן לביצוע…
@omerNLP @clu_avi @mataneyal1 @kroscoo @rtsarfaty A tokenization paper without @yuvalpi are you mad?!