ServiceNow Research @ServiceNowRSRCH
Unlock work experiences of the future. Join @ServiceNowRSRCH as we advance the state-of-the-art in Enterprise AI. #ServiceNowResearch #LifeAtNow #Hiring servicenow.com/research/ Montréal, Québec Joined July 2016-
Tweets2K
-
Followers18K
-
Following1K
-
Likes4K
Conditioning your LLM on merely the last decoder-layer representations of the context is enough for QA with autoregressive models. And you can cache these representations for effective inference! Our paper explains how. arxiv.org/abs/2404.15420 @ServiceNowRSRCH
Help us improve the evaluation harness, by adding more benchmarks and features ✨ Repo: github.com/bigcode-projec…
One of most intriguing findings of 2023 is that adversarial triggers that jailbreak one or more LLMs transfer to other models. We were so excited that we spent many months figuring out the conditions for universal transfer but the transfer never happened. It wasn't a bug 😀
One of most intriguing findings of 2023 is that adversarial triggers that jailbreak one or more LLMs transfer to other models. We were so excited that we spent many months figuring out the conditions for universal transfer but the transfer never happened. It wasn't a bug 😀
🧵) We unexpectedly reach 🥇 on the leaderboard of #WebArena. While 25% is still far from human performance it is a large jump compared to the next best result. The performance gain is largely attributed to #BrowserGym github.com/ServiceNow/Bro… leaderboard: bit.ly/3QjOL5r
Hiring a staff applied researcher in the Trust & Gov Lab at @ServiceNowRSRCH. Help us crack hard problems in genAI risk detection, model eval & governance for risk management, safety, security. Lots of hard, emerging challenges to sink your teeth into! jobs.smartrecruiters.com/ServiceNow/743…
We introduce LLM2Vec, a simple approach to transform any decoder-only LLM into a text encoder. We achieve SOTA performance on MTEB in the unsupervised and supervised category (among the models trained only on publicly available data). 🧵1/N Paper: arxiv.org/abs/2404.05961
How capable are web agents at solving knowledge work tasks? 🤔 Are LLMs up to the challenge? 🤖 Introducing WorkArena: a benchmark where agents meet the world 𝘸𝘪𝘭𝘥 web of enterprise software 🌐🖥️ Paper: bit.ly/4a7FiFV Website: bit.ly/3VkdJ87 🧵 1/7
Having been deeply involved in shaping and driving open innovation over the years, I was asked to share my thoughts on the role and importance of bias towards open vs. closed AI innovation… spoiler alert, we need more open AI innovation!! Read why: servicenow.com/workflow/custo…
Very excited to present a keynote at GTC. If you're around join at ballroom 4.
Very excited to present a keynote at GTC. If you're around join at ballroom 4.
We’re excited to share our latest research publication and introduce you to WorkArena. Learn more in our thread below. #EnterpriseAI #LLM #Benchmark #Automation #FutureOfWork
We’re excited to share our latest research publication and introduce you to WorkArena. Learn more in our thread below. #EnterpriseAI #LLM #Benchmark #Automation #FutureOfWork
My video on our EMNLP2023 paper "PromptMix" is finally on youtube. youtube.com/watch?v=xhrq3c…. It is in collaboration with my ServiceNow & University of Waterloo team where the idea is to use LLMs to perform state-of-the-art data augmentation for text. Hope you enjoy it.
Congratulations to the @BigCodeProject and @SWHeritage communities on developing and releasing the #StarCoder2 code LLM foundation models and The Stack v2 dataset, and to our research partners @huggingface and @nvidia for training the models. #OpenScience #OpenSource #FTW
Congratulations to the @BigCodeProject and @SWHeritage communities on developing and releasing the #StarCoder2 code LLM foundation models and The Stack v2 dataset, and to our research partners @huggingface and @nvidia for training the models. #OpenScience #OpenSource #FTW
Thrilled to share the release of StarCoder2! @ServiceNow , @huggingface, and @nvidia have partnered to deliver a family of open-access code LLMs to help developers everywhere tap the power of GenAI to build software better. Check out model checkpoints on the Hugging Face Hub!
Thrilled to share the release of StarCoder2! @ServiceNow , @huggingface, and @nvidia have partnered to deliver a family of open-access code LLMs to help developers everywhere tap the power of GenAI to build software better. Check out model checkpoints on the Hugging Face Hub!
Foundation models are well-established in vision and language, but time series forecasting has lagged behind - it still relies on dataset-specific models. Meet Lag-Llama: the first open-source foundation model for time series forecasting!
🚀 Seeking Student Researcher Are you interested in LLM-based Agents that can extract insights from data? 🤖📊 We are expanding on our research work [1] to create a comprehensive capture-the-flag benchmark. Apply here: forms.gle/8KfDSUTqLLYsop… [1] openreview.net/forum?id=fkz5V…
🔔 Excited to share that the blog post for our #EMNLP2023 paper, PromptMix: A Class Boundary Augmentation Method for Large Language Model Distillation, is now live! Blog: twtr.to/Vx_y7 Paper: twtr.to/OvH4M
🔔 Excited to share that the blog post for our #EMNLP2023 paper, PromptMix: A Class Boundary Augmentation Method for Large Language Model Distillation, is now live! Blog: twtr.to/Vx_y7 Paper: twtr.to/OvH4M
Kyunghyun Cho @kchonyc
61K Followers 2K Following a combination of a mediocre scientist, a mediocre manager, a mediocre advisor & a mediocre PC at @nyuniversity (@CILVRatNYU) & @genentech (@PrescientDesign).Delip Rao e/σ @deliprao
46K Followers 5K Following Busy inventing the shipwreck. @Penn. Past: @johnshopkins, @UCSC, @Amazon, @Twitter ||Art: #NLProc, Vision, Speech, #DeepLearning || Life: 道元, improv, running 🌈Dan Roy @roydanroy
45K Followers 2K Following ML / AI researcher, emphasis on theory. Research Director and Canada CIFAR AI Chair, @VectorInst Professor, @UofT (Statistics/CS)Eric Jang @ericjang11
69K Followers 3K Following physical AGI at 1X. Author of "AI is Good for You" https://t.co/eFg4WXhg0pIrina Rish @irinarish
9K Followers 994 Following prof UdeM/Mila; Canada Excellence Research Chair; AAI Lab head https://t.co/UzlrC7ZrGF; INCITE project PI https://t.co/0rV7szd7rH; CSO https://t.co/XDhj6MEtUjPablo Samuel Castro @pcastr
10K Followers 813 Following Señor swesearcher @ Google DeepMind. Adjunct prof @ U de Montreal & Mila. Musician. From 🇪🇨 living in 🇨🇦.Natasha Jaques @natashajaques
25K Followers 1K Following Senior Research Scientist at @GoogleAI and Assistant Professor @uwcse. Social Reinforcement Learning in multi-agent and human-AI interactions. PhD from @MIT.Sander Dieleman @sedielem
50K Followers 2K Following Research Scientist at Google DeepMind. I tweet about deep learning (research + software), music, generative models (personal account).Michael Bronstein @mmbronstein
43K Followers 4K Following #DeepMind Professor of #AI @UniofOxford / Fellow @ExeterCollegeOx / ML Lead @ProjectCETI / https://t.co/kZpGpDzYeVHattie Zhou @oh_that_hat
5K Followers 765 Following Finding \hat{y} Give me anonymous feedback: https://t.co/7aBNrpbad8Thomas Wolf @Thom_Wolf
68K Followers 4K Following Co-founder and CSO @HuggingFace - open-source and open-scienceMilaQuebec @Mila_Quebec
31K Followers 561 Following The world's largest academic research center in deep learning — Le plus grand centre de recherche universitaire en apprentissage profond.Ethan Caballero is bu.. @ethanCaballero
8K Followers 2K Following ML PhD student @Mila_Quebec ; previously @GoogleDeepMindTaco Cohen @TacoCohen
21K Followers 3K Following Deep learner at FAIR. Into codegen, equivariance, generative models. Spent time at Qualcomm, Scyfer (acquired), UvA, Deepmind, OpenAI.Andreas Kirsch 🇮�.. @BlackHC
9K Followers 5K Following Past: 🧑🎓 DPhil @AIMS_oxford @ExeterCollegeOx @UniofOxford (4.5yr) 🧙♂️ RE @DeepMind (1yr) 📺 SWE @Google (3yrs) 🎓 @TU_Muenchen 👤 Fellow @nwspkLeo Boytsov @srchvrs
7K Followers 2K Following Sr. Research Scientist @AWS Labs (ph-D @LTIatCMU) working on unnatural language processing, speaking πtorch & C++. Opinions sampled from MY OWN 100T param LM.Shane Gu @shaneguML
28K Followers 1K Following Research Scientist & Manager @GoogleDeepMind Tokyo/MTV. ex: @GoogleAI Brain, @OpenAI. (JP: @shanegJP)Guillaume Lajoie @g_lajoie_
2K Followers 235 Following Associate Prof. @UMontreal Math & Stats and @MILAMontreal, @ai_unique Applied Math, AI, Computational Neuroscience, Neural InterfacingBigCode @BigCodeProject
9K Followers 3 Following Open and responsible research and development of large language models for code. #BigCodeProject run by @huggingface + @ServiceNowRSRCHKBLi @KbAr0cLi
11 Followers 341 Following@VagabondSA @VagabondguySA
173 Followers 334 Following Here for positive change in favour of humanity.Akk Shar @Akk_Shar
88 Followers 130 Following Believe in Nation 🇮🇳 First... भारत माता की जय....🇮🇳....राष्ट्र सर्वप्रथम सर्वोपरि....🇮🇳....🕉️Veri @missmanonve
44 Followers 214 FollowingKrushnakumar Thombare @Krushnakumar_15
3 Followers 104 FollowingVaibhav Phalke @PhalkeVaibhav07
0 Followers 9 FollowingOzair Saiyad @ozairsfdc
1 Followers 16 Following Documenting my journey to becoming a Servicenow developer, and sharing my learnings along the wayBashir @Bashir710158
5 Followers 91 Followingいっと @E3jiDPNtDS6046
13 Followers 166 FollowingZizou @ifukyourmok
14 Followers 23 FollowingKellie Strobel @KStrobel47412
1 Followers 4 FollowingCarsten Lindstedt @clindstedt
583 Followers 3K Following Creativity ✗ Technology; Head of Brand Experience & Gen AI evangelist @Dassault3DS 3DEXCITE, ex Creative Director UX/UI & startup co-founder (acquired by Bosch)Soham Parikh @sohampar
0 Followers 7 FollowingHarry Kong @harrystoneocean
244 Followers 5K Following Program Manager in Tech / Wandering around the worldAnkur Bohra @AnkurBohra9
19 Followers 2K FollowingBlueprint Solutions @blueprintsols
0 Followers 5 Following Specializing in application development and educational services, dedicated to assisting ServiceNow platform owners and admins to hit the ground running.kindle @kindle_jiang
20 Followers 416 FollowingAvinash @avilearning
1 Followers 5K FollowingJan "yawn" 🥱🥲 @yawnxyz
2K Followers 4K Following Tech, AI, biotech. Full-stack UX engineer. @phagedirectory. ex @ubiquiti @phageaustralia CMU · ⚔️ 🦠 🇸🇪 🇺🇸 🇦🇺Jean Defayolle @defayolle
8 Followers 244 FollowingGabriel @GabrielDekris
0 Followers 548 FollowingGreen Clouds @_greenclouds
51 Followers 125 FollowingEldon Brown @eldonbrown
89 Followers 106 FollowingRMO Digital @RMODigital
98 Followers 2K Followingprettypissies @4prettypissies
9 Followers 55 FollowingKathryn Hill @KathrynHil24557
1 Followers 22 FollowingRaul Pesch @raulpesch
1K Followers 3K FollowingSHOCKWAVAI @shockwavAI
0 Followers 4 FollowingArthur Nshimirimana @ArthurNshi74501
1 Followers 49 FollowingRayan H. Assaad @RayanAssaad
145 Followers 1K Following Assistant Professor of Civil Engineering; New Jersey Institute of Technology (NJIT)楼明 @lumng125194
8 Followers 84 FollowingEl🌵 @_elakiya_
47 Followers 166 Following With freedom, books, flowers and the moon, who could not be happy with?Triumph Wrestling @TriumphRoy
678 Followers 233 Following Triumph Wrestling will push you to the limit so you can reach your ultimate potential. Get triumph trainedDinaliza™ 🇪🇬�.. @dinaibrahim0
556 Followers 1K Following #ComputerScience #FCIS #ASU #Egypt #Scorpio أذكروا الله - لعله خير - نيتي صافيةEdward de Minckwitz @eddeminckwitz
77 Followers 1K FollowingSamuel Vilkovský @VilkovskyS
7 Followers 117 Following Servicenow Developer @ Siemens Healthineers /CSA, CAD - certified 🤓/ & UNI AI Student 🤖Bryon Acton 🇮🇱 @BryonActon
195 Followers 2K Following I hope you have not been leading a double life, pretending to be wicked and being good all the time. That would be hypocrisy.Soumith Chintala @soumithchintala
187K Followers 883 Following Cofounded and lead @PyTorch at Meta. Also dabble in robotics at NYU. AI is delicious when it is accessible and open-source.(((ل()(ل() 'yoav))).. @yoavgo
46K Followers 2K FollowingRichard Socher @RichardSocher
101K Followers 971 Following CEO @youSearchEngine Investing at @aixventuresHQ Before: Stanford Adj Prof in AI/NLP, Chief Scientist at Salesforce, MetaMindDelip Rao e/σ @deliprao
46K Followers 5K Following Busy inventing the shipwreck. @Penn. Past: @johnshopkins, @UCSC, @Amazon, @Twitter ||Art: #NLProc, Vision, Speech, #DeepLearning || Life: 道元, improv, running 🌈Dan Roy @roydanroy
45K Followers 2K Following ML / AI researcher, emphasis on theory. Research Director and Canada CIFAR AI Chair, @VectorInst Professor, @UofT (Statistics/CS)NeurIPS Conference @NeurIPSConf
112K Followers 35 Following New Orleans, Dec 10-16, 23. https://t.co/ga8aOw615g Tweets to this account are not monitored. Please send feedback to [email protected].Miles Brundage @Miles_Brundage
43K Followers 10K Following Policy research at @openai. I mostly tweet about AI, animals, and sci-fi. He/him. Views my own.Natasha Jaques @natashajaques
25K Followers 1K Following Senior Research Scientist at @GoogleAI and Assistant Professor @uwcse. Social Reinforcement Learning in multi-agent and human-AI interactions. PhD from @MIT.Sasha Rush @srush_nlp
52K Followers 464 Following Professor, Programmer in NYC. Cornell Tech, Hugging Face 🤗 https://t.co/cZl0wTfqGzSander Dieleman @sedielem
50K Followers 2K Following Research Scientist at Google DeepMind. I tweet about deep learning (research + software), music, generative models (personal account).Sergey Levine @svlevine
80K Followers 122 Following Associate Professor at UC Berkeley Co-founder, Physical IntelligenceEdward Grefenstette @egrefen
36K Followers 776 Following FR/US/GB AI/ML Person, Director of Research at @GoogleDeepMind, Honorary Professor at @UCL_DARK, @ELLISforEurope Fellow. All posts are personal.MilaQuebec @Mila_Quebec
31K Followers 561 Following The world's largest academic research center in deep learning — Le plus grand centre de recherche universitaire en apprentissage profond.Ethan Caballero is bu.. @ethanCaballero
8K Followers 2K Following ML PhD student @Mila_Quebec ; previously @GoogleDeepMind🇺🇦 Dzmitry Bahd.. @DBahdanau
6K Followers 36 Following Research Scientist & Research Lead at ServiceNow Research Adjunct Prof @ McGill. Member of Mila, Quebec AI Institute. Stream of consciousness is my own.BigCode @BigCodeProject
9K Followers 3 Following Open and responsible research and development of large language models for code. #BigCodeProject run by @huggingface + @ServiceNowRSRCHBlake Richards @tyrell_turing
15K Followers 2K Following Researcher at @mcgillu combining AI and neuroscience. Also on Bluesky (@tyrellturing.bsky.social) and Mastodon: @[email protected].Graham Neubig @gneubig
31K Followers 588 Following Associate professor at CMU, studying natural language processing and machine learning.Zdeněk Kasner @ZdenekKasner
93 Followers 74 Following PhD student and Researcher at Charles University.Abhay Puri @AbhayPuri98
576 Followers 923 Following Visiting Researcher @ServiceNowRSRCH| ex-MLE @Jumio | Grad Student @Mila_QuebecNicolas Gontier @nicogontier
349 Followers 367 Following Research Scientist @ServiceNowRSRCH; PhD from @polymtl and @Mila_QuebecThe Rundown AI @TheRundownAI
131K Followers 100 Following Daily AI newsletter with over 500,000+ readers. Get the latest in AI and learn how to apply it in 5 minutes. By @rowancheungLevi @levikul09
38K Followers 188 Following I explain Data Science on Grandma's level. Writing https://t.co/25jLCDRZmsTobias Zwingmann @ztobi
2K Followers 472 Following Chief AI Augmentor | Author | Speaker - Helping forward-looking B2Bs innovate and grow with AI (no big tech team needed). Featured on CNBC.Harm de Vries @harmdevries77
1K Followers 154 Following Building something new | prev co-lead @BigCodeProject @ServiceNowRsrch | PhD from @Mila_QuebecThe Good Men Project @GoodMenProject
167K Followers 131K Following We're having a conversation about what it means to be a good man. Join us! https://t.co/Dd6zq6Zwbt [email protected]Humanloop @humanloop
7K Followers 438 Following The enterprise platform for developing and evaluating LLM applicationsSean Hughes @hughesthe1st
555 Followers 269 Following AI Ecosystem @ServiceNow @ServiceNowRSRCH @BigCodeProject #TheAIAlliance - formerly @IntelAI @ActianCorp @HPE - All tweets are my own opinion.Loubna Ben Allal @LoubnaBenAllal1
4K Followers 623 Following ML Engineer @huggingface 🤗 | @ENS_ParisSaclay - MVALeandro von Werra @lvwerra
6K Followers 310 Following Machine learning @huggingface: co-lead of @bigcodeproject and maintainer of TRL.Catherine Martin @cathe_martin
107 Followers 362 Following Research Internship Program Manager at @ServiceNowRSRCH. Former Element AI.João Monteiro @joaomonteirof
210 Followers 933 Following Research Scientist @ServiceNowRSRCH. Opinions are my own.Valentina Zantedeschi @vzantedesc
279 Followers 364 Following Sr Research Scientist at @ServiceNowRSRCH, formerly @ai_ucl @Inria_Lille @FDL_Europe @LabHubertCurien @IBMResearchEEML @EEMLcommunity
3K Followers 11 Following Strengthening the Eastern European ML community and improving diversity in the field. https://t.co/34QAbYBeDoAntonin Schrab @AntoninSchrab
401 Followers 453 Following PhD student in Foundational AI at @ai_ucl & @GatsbyUCL. Kernel methods, hypothesis testing, generative models.Sai Rajeswar @RajeswarSai
336 Followers 316 Following Senior Research Scientist @ ServiceNow Research. Previously research @mila_quebec @ Google Deepmind. Views my own.Chuck Tomasi @ctomasi
4K Followers 622 Following Dad, husband, IT geek, podcaster, black belt, author, volunteerLee Zamparo 🖖 @lzamparo
983 Followers 803 Following Ex @recursionpharma, Ex @ServiceNowRSRCH, Ex @sloan_kettering, Ex @UofTCompSci PhD. Leaning into Ex, not X. See you on the blue site, HMU for invites.Pierre-André Noël @PierreAndreNoel
17 Followers 19 Following Applied Research Scientist @ServiceNowRSRCH. All opinions are my own.ServiceNow Events @HelloKnowledge
10K Followers 441 Following Come curious. Leave energized. Join us at a @ServiceNow event. 🗓️ #Know24 May 7-9th ➡️ Register: https://t.co/aaXndhMG8WNow Support @Nowsupport
7K Followers 368 Following The official account for the @ServiceNow Support digital experience.ServiceNow Community @WeUseServiceNow
11K Followers 520 Following We're here to help and cheer on our problem solvers and innovators reinventing how work gets done. 📣👏 Today is a great day to start learning @ServiceNow.Bill McDermott @BillRMcDermott
56K Followers 263 Following Owned a deli as a kid, got my big break @Xerox, transformed @SAP, now leading @ServiceNow. Relentless optimist. Devoted to family, friends & a better world.ServiceNow Germany @ServiceNowDE
1K Followers 1K Following Besser arbeiten mit ServiceNow. Unsere Cloud-basierte Plattform und Lösungen bieten digitale Erlebnisse, die Menschen bei ihrer Arbeit optimal unterstützen.ServiceNow Japan @ServiceNowJapan
2K Followers 138 Following ServiceNowは企業全体の働き方を改善するという考えのもと、シンプルな業務から複雑なタスクに至るまで従業員が簡単に行える仕組みを実現します。ServiceNow France @ServiceNowFR
2K Followers 885 Following ServiceNow simplifie le travail. Notre plateforme & nos solutions cloud offrent des expériences digitales pour se concentrer davantage sur son cœur de métier.ServiceNow Asia @ServiceNowAsia
2K Followers 415 Following ServiceNow makes work, work better for people. Our cloud-based platform and solutions deliver digital experiences that help people do their best work.ServiceNow ANZ @ServiceNowANZ
2K Followers 440 Following The world works with ServiceNow. Our cloud-based platform and solutions deliver digital experiences that help people do their best work.ServiceNow UK and Ire.. @ServiceNowUKI
7K Followers 1K Following ServiceNow makes work, work better for people. Our cloud-based platform and solutions deliver digital experiences that help people do their best work.Hector Palacios @hectorpal
2K Followers 2K Following Research Scientist at @ServiceNowRSRCH. We are hiring! I’m interested in AI, politics, sociology, philosophy, art. That’s like Artificial BehaviourDave Wright @TheWrightView
3K Followers 368 Following Driving strategy at ServiceNow - IoT, user experience, AI and what's next for cloud.... old enough to know what TSO stands for - STILL buying stupid cars.Deep Learning For Cod.. @DL4Code
482 Followers 23 Following ICLR 2023 Deep Learning For Code (DL4C) WorkshopVijay K Narayanan @VijayKNarayana1
203 Followers 437 Following Data for Good. Day job: SVP Dev Engineering | Chief AI Officer at ServiceNowLife at ServiceNow @LifeAtNow
6K Followers 444 Following At @ServiceNow, our work makes the world work. Join us! #LifeAtNowICLR 2024 @iclr_conf
41K Followers 40 Following International Conference on Learning Representations #ICLR2024. SPC is @yisongyue and GC is @_beenkim OpenReview:https://t.co/OD1sg0r7F8ServiceNow News @ServiceNowNews
4K Followers 188 Following The world works with @ServiceNow. Follow us here for what's new and noteworthy. Please send media requests to [email protected].Torsten Scholak @tscholak
2K Followers 3K Following research scientist @ServiceNowRSRCH working on functional deep learning, program synthesis, and semantic parsing. opinions are not that of my employerServiceNow @ServiceNow
49K Followers 194 Following The world works with ServiceNow @ServiceNowNews | @WeUseServiceNow | @LifeAtNow | @HelloKnowledgeManon Gruaz ❤️.. @manongruaz
847 Followers 2K Following ♥︎ Design Lead @HaleoClinic • I Heal by Design • MentalHealth advocate • future Keynote speaker ♥︎ focus: #UX #design, #mentalhealth and #compassionLayla El Asri @elasri_layla
2K Followers 836 Following Senior machine learning research team lead working with the amazing team at Borealis AI. Pronouns are she/her. Views are my own. Currently on a Twitter hiatusMassimo Caccia @MassCaccia
2K Followers 569 Following Research Scientist @ServiceNowRSRCH. Gradient-descent enthusiast building LLM agents. PhD @Mila_Quebec. Formerly @Deepmind, @AWScloud, @SpotifyResearch.MONTREAL.AI @Montreal_AI
179K Followers 178 Following https://t.co/LziAg5YHcG | AGI.Eth #AGI Company 🗝️👾 Disc.: https://t.co/FchXRDk4ki OS: https://t.co/dbjdfMoPxb Français: @Montreal_IA #MontrealAIReleasing 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…
The age of AI is now. It's not about *if* but *how* to put AI to work for your business. Join CEO Bill McDermott's #Know24 opening keynote as he leads a conversation around building the future of business on the ServiceNow platform. 🚀 See you there! spr.ly/61105ZQepv
@oakela at @dltHub showed how to fine-tune StarCoder 2 using Continue tab-autocomplete data at the @ollama open source AI meetup yesterday @joinstationf🦙🇫🇷
Open source LLMs popping up everywhere. This is the way.
Today is the Llama moment for coding! 💫 StarCoder-15B reaches 40.8% on HumanEval benchmark, beating the 30x bigger PaLM. Coding holds a very special place in NLP. Most software in the world has AI-friendly APIs. LLMs good at coding will master the digital tools, greatly…
@julien_c The authors missed the chance to make the spaces represent a hat
.@ncmeade has built wonderful code repo to run GCG optimization faster to find adversarial triggers. Please check it out github.com/McGill-NLP/adv…
One of most intriguing findings of 2023 is that adversarial triggers that jailbreak one or more LLMs transfer to other models. We were so excited that we spent many months figuring out the conditions for universal transfer but the transfer never happened. It wasn't a bug 😀
Adversarial Triggers For LLMs Are 𝗡𝗢𝗧 𝗨𝗻𝗶𝘃𝗲𝗿𝘀𝗮𝗹!😲 It is believed that adversarial triggers that jailbreak a model transfer universally to other models. But we show triggers don't reliably transfer, especially to RLHF/DPO models. Paper: arxiv.org/abs/2404.16020
🚀 LLAMA3 is the first open-source LLM to ace tasks in Workarena, making it the top OSS virtual knowledge worker! (and believe us we've tested many models and prompting techniques 😅) Watch it excel in a challenging knowledge base task. Kudos to @AIatMeta for the amazing model 🎉
Very excited about #Llama3 . We are finally able to get some tasks to be solved in #WorkArena using open-source LLMs. Previous attempts always gave 0% except for GPT. servicenow.github.io/WorkArena/
🚀 LLAMA3 is the first open-source LLM to ace tasks in Workarena, making it the top OSS virtual knowledge worker! (and believe us we've tested many models and prompting techniques 😅) Watch it excel in a challenging knowledge base task. Kudos to @AIatMeta for the amazing model 🎉
@MassCaccia @ServiceNowRSRCH @Mila_Quebec Excited about it too!
Mistral is not confused when we enable bidirectionality whereas LLaMA goes off the rails 🤠. We may have unlocked one secret ingredient of why Mistral is better than LLaMA. We believe it is 💥Prefix LM💥. This side finding is exciting in itself!
We also analyze how enabling bidirectional attention without training affects the representations of decoder-only LLMs 🔍. We find that Mistral-7B is surprisingly good at using bidirectional attention out-of-the-box 🤯 and speculate it was likely trained as a prefix-LM 🤔. 7/N
We introduce LLM2Vec, a simple approach to transform any decoder-only LLM into a text encoder. We achieve SOTA performance on MTEB in the unsupervised and supervised category (among the models trained only on publicly available data). 🧵1/N Paper: arxiv.org/abs/2404.05961
It's easy to see the value of open science with a team of @BigCodeProject's quality, with such strong leaders as @harmdevries77, @lvwerra and @hughesthe1st !
@lvwerra @ServiceNow 💯 ! A big thanks to @NicolasChapados for understanding the value of open-science for businesses like ServiceNow
Open innovation is ‘the rising tide that floats all boats’ - at ServiceNow we take a hybrid approach to innovation in generative AI, supporting both fundamental open scientific research and in creating customer value through applied AI/ML as we develop advanced technologies for…
Why would any company release a strong LLM for free? This is @ServiceNow (mkt cap ~$160Bn) stock price since the release of StarCoder. Only required a small amount of compute and a handful of people while building up a lot of valuable know-how fast. It's not a zero-sum game!
@lvwerra @ServiceNow 💯 ! A big thanks to @NicolasChapados for understanding the value of open-science for businesses like ServiceNow
Thrilled to introduce WorkArena, the first benchmark to evaluate web agents on real-world knowledge work tasks, on a real-world environment!
How capable are web agents at solving knowledge work tasks? 🤔 Are LLMs up to the challenge? 🤖 Introducing WorkArena: a benchmark where agents meet the world 𝘸𝘪𝘭𝘥 web of enterprise software 🌐🖥️ Paper: bit.ly/4a7FiFV Website: bit.ly/3VkdJ87 🧵 1/7
How capable are web agents at solving knowledge work tasks? 🤔 Are LLMs up to the challenge? 🤖 Introducing WorkArena: a benchmark where agents meet the world 𝘸𝘪𝘭𝘥 web of enterprise software 🌐🖥️ Paper: bit.ly/4a7FiFV Website: bit.ly/3VkdJ87 🧵 1/7
WorkArena is built on the ServiceNow platform, which is used by most of the Fortune 500 companies. It contains 29 tasks, which encompass the most common ways workers interact with the software. For example, searching for information in a company knowledge base ⬇️ 2/7
Developing enterprise web-based agents is no small feat. Challenges include: 🌐 Navigating the 'World 𝘞𝘪𝘭𝘥 Web' of Enterprise UI with huge and non-standard HTML files (100k+ tokens) 💾 Memory management (for POMPDs) 📈 Planning 📚 Domain-specific knowledge 3/7
To help agents navigate this space we propose a new web environment with: Multimodal observations (AXtree, HTML, screenshots) ☑High-level actions (click, fill, etc.) ☑Chat interface for user interaction ☑Robustness to the world wild web (e.g., iFrames, shadow DOMs) 4/7