Jacob Steinhardt @JacobSteinhardt
Assistant Professor of Statistics, UC Berkeley Joined December 2011-
Tweets322
-
Followers7K
-
Following67
-
Likes85
In a new preprint with Jarek Blasiok, @rares_buhai, David Steurer, we show a surprisingly simple greedy algorithm that can list decode planted cliques in the semirandom model at k~sqrt n log^2 n --essentially optimal up to log^2 n. This ~resolves @JacobSteinhardt's open question.
In a new preprint with Jarek Blasiok, @rares_buhai, David Steurer, we show a surprisingly simple greedy algorithm that can list decode planted cliques in the semirandom model at k~sqrt n log^2 n --essentially optimal up to log^2 n. This ~resolves @JacobSteinhardt's open question.
Super excited to share that VisDiff has been accepted to #CVPR2024 and selected as an oral (90/11,532)! We will give a 15-min presentation going through the methods and exciting applications enabled by VisDiff. See you in Seattle!
Super excited to share that VisDiff has been accepted to #CVPR2024 and selected as an oral (90/11,532)! We will give a 15-min presentation going through the methods and exciting applications enabled by VisDiff. See you in Seattle!
Language models can imitate patterns in prompts. But this can lead them to reproduce inaccurate information if present in the context. Our work (arxiv.org/abs/2307.09476) shows that when given incorrect demonstrations for classification tasks, models first compute the correct…
Protein language models (pLMs) can give protein sequences likelihood scores, which are commonly used as a proxy for fitness in protein engineering. But what do likelihoods encode? In a new paper (w/ @JacobSteinhardt) we find that pLM likelihoods have a strong species bias! 1/
Independent AI research should be valued and protected. In an open letter signed by over a 100 researchers, journalists, and advocates, we explain how AI companies should support it going forward. sites.mit.edu/ai-safe-harbor/ 1/
Beating prediction markets with chatbots sounds cool. In a recent work arxiv.org/abs/2402.18563, we get somewhat close to that. As another perspective, forecasting is a great capability domain to benchmark LM reasoning, calibration, pre-training knowledge, and more. 🧵1/n
Remember when @bing’s LLM Sydney threatened @marvinvonhagen for tweeting about its prompt? Our paper shows how such unexpected behavior in LLMs emerges from feedback loops and provides recommendations for evaluation to capture feedback effects. 📰: arxiv.org/abs/2402.06627 1/
Accepted to oral #ICLR2024! *Interpreting CLIP's Image Representation via Text-Based Decomposition* CLIP produces image representations that are useful for various downstream tasks. But what information is actually encoded in these representations? [1/8]
Percy Liang @percyliang
49K Followers 408 Following Associate Professor in computer science @Stanford @StanfordHAI @StanfordCRFM @StanfordAILab @stanfordnlp | cofounder @togethercompute | Pianist
Yi Ma @YiMaTweets
71K Followers 123 Following Chair Professor in AI, Director of IDS, Head of CS, HKU; Professor of EECS, Berkeley; Author of Book: High-Dim Data Analysis, https://t.co/gwaqMJp8av.
Neel Nanda @NeelNanda5
13K Followers 89 Following Mechanistic Interpretability lead @DeepMind. Formerly @AnthropicAI, independent. In this to reduce AI X-risk. Neural networks can be understood, let's go do it!
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)
Jelani Nelson @minilek
22K Followers 184 Following Professor @Berkeley_EECS. Research Scientist (part-time) @GoogleAI. Founder @addiscoder. 🇻🇮🇺🇸🇪🇹
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.
Jan Leike @janleike
44K Followers 322 Following ML Researcher, co-leading Superalignment @OpenAI. Optimizing for a post-AGI future where humanity flourishes.
Sam Bowman @sleepinyourhat
35K Followers 3K Following AI alignment + LLMs at NYU & Anthropic. Views not employers'. No relation to @s8mb. I think you should join @givingwhatwecan.
David Krueger @DavidSKrueger
13K Followers 4K Following Cambridge faculty - AI alignment, deep learning, and existential safety. Formerly Mila, FHI, DeepMind, ElementAI, AISI.
Jacob Andreas @jacobandreas
14K Followers 958 Following Teaching computers to read. Assoc. prof @MITEECS / @MIT_CSAIL (he/him). https://t.co/5kCnXHjtlY https://t.co/2A3qF5vdJw
Behnam Neyshabur @bneyshabur
18K Followers 690 Following Senior Staff Research Scientist @GoogleDeepMind, Interested in reasoning w. LLMs, traveling & backpacking
Amanda Askell @AmandaAskell
26K Followers 653 Following Philosopher & ethicist teaching models to be good @AnthropicAI. Personal account. All opinions come from my training data.
Boaz Barak @boazbaraktcs
17K Followers 422 Following Computer Scientist. See also https://t.co/EXWR5k634w, https://t.co/SEVX6it6z3 ( @[email protected] , boaz.barak in threads ). Opinions my own.
Peter Wildeford @peterwildeford
10K Followers 367 Following Pro forecaster w/ good track record. Seeking to understand + manage risks from advanced AI systems. - Co-CEO @RethinkPriors - Chief Advisory Executive @iapsAI
Horace He @cHHillee
23K Followers 449 Following Working at the intersection of ML and Systems @ PyTorch "My learning style is Horace twitter threads" - @typedfemale
Ethan Caballero is bu.. @ethanCaballero
8K Followers 2K Following ML PhD student @Mila_Quebec ; previously @GoogleDeepMind
Zachary Lipton @zacharylipton
59K Followers 2K Following Professor: CMU/@acmi_lab, CTO / CSO: @AbridgeHQ, Creator: @d2l_ai & https://t.co/QQt98VNLUp, Relapsing 🎷
Catherine Olsson @catherineols
15K Followers 1K Following Hanging out with Claude, improving its behavior, and building tools to support that @AnthropicAI 😁 prev: @open_phil @googlebrain @openai (@microcovid)
Collin Burns @CollinBurns4
11K Followers 276 Following Superalignment @OpenAI. Formerly @berkeley_ai @Columbia. Former Rubik's Cube world record holder.
125xzc @125xzc41237
1 Followers 13 Following
Open @OpenXuu
0 Followers 150 Following
ante @ante33277969
14 Followers 87 Following
Claudia Richoux @_laudiacay
2K Followers 341 Following @banyancomputer is decentralizing the cloud // ex @protocollabs @trailofbits @uchicago
Yijun Dong @YijunDong1
49 Followers 146 Following Postdoc at NYU Courant. PhD from UT Austin. Interested in randomized numerical linear algebra and machine learning.
Nikita @nikitavoloboev
4K Followers 7K Following Make @LearnAnything_ Learn in public: https://t.co/GbFvuErkYn macOS course: https://t.co/JdbJWru6zG https://t.co/94R8ER7K2h https://t.co/ROkqhyhpEK
Ian @ InfoHunt.ai @Ianyan2023
34 Followers 231 Following [email protected],Your Most Reliable Discovery AI Engine 👉 Click to explore: https://t.co/WkjTFNHdCr
Yuchen Zhu @_zhuyuchen
288 Followers 276 Following PhD candidate in Causal Inference @UCL. Interested in causal inference and abstraction. Ex: @MSFTResearchCam, @amazon, Mathematics @Cambridge_Uni.
Sonakshi Chauhan @ChauhanSon8200
12 Followers 36 Following
Constantin Venhoff @cvenhoff00
3 Followers 19 Following
Rais Latif @RaisLatif_Study
39 Followers 5K Following Hi I'm Rais. I'm mainly focussing on Math and Science lifelong. There is a lot to discover in these fields and my mind is always blown by all the cool things.
whimchic @whimchic
3K Followers 5K Following *Fashion, Tech 🟣, Culture & Style. #betatester. #contentcreator. #aichatter-er.
Gautham Elango @gautham_elango
635 Followers 2K Following
Gabriel Simmons @gs1mm0ns
43 Followers 422 Following
YuvaRaj @YuvaAnandan
2 Followers 37 Following Enthusiast, angle investor, worked as Change agent in Finance services and currently working with Payments fintech as Product Lead
Raj Contractor @RajContrac26606
1 Followers 40 Following
Rain @RainyNacht
25 Followers 124 Following A grad student. Interested in graph algorithms, combinatorial optimization and parameterized complexity.
coffee & AI @realcoffeeAI
51 Followers 721 Following Sitting on a park bench scattering random seeds for the LLMs. I never bet against Elon.
Alex Wozniakowski @airwoz
164 Followers 157 Following
Nikos Karalias @AspectStalence
348 Followers 1K Following Postdoc at MIT CSAIL. Working on combinatorial optimization with neural nets https://t.co/bf4UWg2BUQ
Aurora @Yuchichuan123
3K Followers 2K Following San Francisco Food + Travel 💁🏼♀️ 💎 Local view of San Francisco 🏠 A unique SF trip home; Eating. 🚧 Similar hobby 👩🏼⚕️ Full time at work
Yatin Dandi @DandiYatin
250 Followers 2K Following PhD student (EPFL, Switzerland), interested in the theory of deep learning and statistical physics of computation
Henrietta.SolidGoldMa.. @Bearly_Present
181 Followers 3K Following All integers between 0 and 1 exclusive.
Keertana VC @KeertanaVc15354
0 Followers 7 Following
Ahmed Yousef @mennanour12
37 Followers 334 Following
Erik Larsson @notahegelbot
25 Followers 480 Following just a phenomenal self-model in a world-model (both subject to calibration)
Meng Li @limengnlp
14 Followers 515 Following PhD student @unipotsdam, supervised by @davidschlangen. Working on NLP, ML and CogSci. Prev @LstSaar. Former NLP engineer.
Bergen & Associates @BergenandAssoc
19 Followers 265 Following
Algeia @Algeia17812
21 Followers 68 Following
Junjie (Jorji) Chen @coderchen01
2 Followers 282 Following
Kunvar Thaman @firstuserhere
220 Followers 640 Following Taking apart neural networks and putting them back together for a living Social profiles: https://t.co/OxoeMvCw3a
Vesa Haapalahti @vjhaapalahti
1K Followers 2K Following Uteliaisuuden kohteena #kauppa #johtaminen, #CX, #AI, #tulevaisuus ja #parempiarki - kotona ja töissä. Country Digital Manager, @IKEASuomi. Twiitit omia.
Ayda Sultan @ayda_hassen
1 Followers 23 Following
Dhruv Trehan @dhruvtrehan9
832 Followers 4K Following Learning to learn and do. He/Him. DMs are open. Prev @metaformsai @stoaHQ @TheCitizen_in
Nitasha Tiku @nitashatiku
68K Followers 8K Following Tech culture reporter @washingtonpost in SF [email protected], Signal: 917-318-7531 (she/her) @ https://t.co/APXCymgGYD @[email protected]
Larissa Schiavo @lfschiavo
1K Followers 1K Following 🤖,💻,🐈⬛,🌱,🎞️// 🇧🇷 - 🇺🇸 // previously @OpenAI @mural @USC // writes
Daniel Tenreiro @TenreiroDaniel
9K Followers 1K Following now: having conversations about AI. before: Thiel Macro, @NRO, @Yale. opinions belong to those who love them.
Ivelina Petrova @ivelinapetrovaX
39 Followers 785 Following Industrial Management and development Master Degree and Architect in Architecture [email protected] [email protected] [email protected]
Kaiying Hou @kaiyinghou
9 Followers 135 Following
Percy Liang @percyliang
49K Followers 408 Following Associate Professor in computer science @Stanford @StanfordHAI @StanfordCRFM @StanfordAILab @stanfordnlp | cofounder @togethercompute | Pianist
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)
Sasha Rush @srush_nlp
52K Followers 464 Following Professor, Programmer in NYC. Cornell Tech, Hugging Face 🤗 https://t.co/cZl0wTfqGz
Sam Bowman @sleepinyourhat
35K Followers 3K Following AI alignment + LLMs at NYU & Anthropic. Views not employers'. No relation to @s8mb. I think you should join @givingwhatwecan.
Jacob Andreas @jacobandreas
14K Followers 958 Following Teaching computers to read. Assoc. prof @MITEECS / @MIT_CSAIL (he/him). https://t.co/5kCnXHjtlY https://t.co/2A3qF5vdJw
Behnam Neyshabur @bneyshabur
18K Followers 690 Following Senior Staff Research Scientist @GoogleDeepMind, Interested in reasoning w. LLMs, traveling & backpacking
Collin Burns @CollinBurns4
11K Followers 276 Following Superalignment @OpenAI. Formerly @berkeley_ai @Columbia. Former Rubik's Cube world record holder.
Kayo Yin @kayo_yin
8K Followers 560 Following PhD student @berkeley_ai @berkeleynlp working on interpretability and signed languages. Former @msftresearch @deepmind @carnegiemellon @polytechnique. 🇫🇷🇯🇵
Dan Hendrycks @DanHendrycks
17K Followers 81 Following • Director of the Center for AI Safety (https://t.co/ahs3LYCpqv) • GELU/ImageNet-C/MMLU/safety groundwork • PhD in AI from UC Berkeley https://t.co/rgXHAnYAsQ https://t.co/YtGtDh1aAV
Owain Evans @OwainEvans_UK
7K Followers 241 Following Research Associate @fhioxford, Oxford University. AI alignment. Prefer email to DM.
Jason Lee @jasondeanlee
10K Followers 3K Following Associate Professor at Princeton and Research Scientist at Google DeepMind. ML/AI Researcher working on foundations of LLMs and deep learning
Danny Halawi @dannyhalawi15
168 Followers 290 Following masters student at @berkeley_ai advised by @JacobSteinhardt. Interested in interpretability, scalable oversight, and forecasting.
Nora Belrose @norabelrose
8K Followers 124 Following Working toward a free and fair future powered by friendly AI. Head of interpretability research at @AiEleuther, but tweets are my own views, not Eleuther’s.
Cas (Stephen Casper) @StephenLCasper
3K Followers 1K Following #AI safety & responsibility. PhD Candidate @ #MIT_CSAIL.
Diyi Yang @Diyi_Yang
14K Followers 2K Following Assistant Professor @Stanford CS @StanfordNLP @StanfordAILab. Formerly @GeorgiaTech. Computational Social Science & NLP
xuan (ɕɥɛn / sh-ye.. @xuanalogue
5K Followers 975 Following PhD Student. MIT ProbComp / CoCoSci. Inverting Bayesian models of human reasoning and decision-making. Pronouns: 祂/伊 Mastodon: @[email protected]
Center for AI Safety @ai_risks
5K Followers 1 Following Reducing societal-scale risks from artificial intelligence through technical research and field-building.
David Duvenaud @DavidDuvenaud
28K Followers 3K Following Machine learning prof @UofT. Working on generative models, inference, & latent structure.
Yejin Choi @YejinChoinka
19K Followers 330 Following professor at UW, director at AI2, adventurer at heart
Chelsea Finn @chelseabfinn
69K Followers 384 Following Asst Prof of CS & EE @Stanford. PhD from @Berkeley_EECS, EECS BS from @MIT
Ted Gibson, Language .. @LanguageMIT
12K Followers 1K Following I am Ted Gibson and I run a language lab at MIT. I tweet about psycholinguistics, cognitive science, language research, linguistics, and words.
kipply @kipperrii
8K Followers 825 Following "drop the forest nymph act we know how much gdp you generate" - @mnovendstern | alt @kipperriiii
Zico Kolter @zicokolter
15K Followers 499 Following Associate professor at Carnegie Mellon, VP and Chief Scientist at Bosch Center for AI. Researching (deep) machine learning, robustness, implicit layers.
Aleksander Madry @aleks_madry
31K Followers 166 Following Head of Preparedness at OpenAI and MIT faculty (on leave). Working on making AI more reliable and safe, as well as on AI having a positive impact on society.
Julius Adebayo @juliusadml
2K Followers 989 Following Building interpretable foundation models that can explain their reasoning and tools to align and them @guidelabsai.
Frances Ding @FrancesDing
241 Followers 103 Following PhD student in EECS, UC Berkeley. ML fairness and interpretability. ML for protein design.
Open Philanthropy @open_phil
15K Followers 17 Following Open Philanthropy's mission is to help others as much as we can with the resources available to us.
Paul Romer @paulmromer
94K Followers 506 Following I no longer post to this account. There is at least one person who is pretending to be me with a handle that adds an "s" to the end of my handle.
ML Safety @ml_safety
1K Followers 2 Following Course: https://t.co/XWcOJjXRVG Newsletter: https://t.co/HEe7NatYhA Papers as they come out: https://t.co/d7f799Sby2. More: https://t.co/NgGDTW4sYk
Robert Nishihara @robertnishihara
6K Followers 623 Following Co-founder and CEO @anyscalecompute. Co-creator of @raydistributed. Previously PhD ML at Berkeley.
Chelsea Sierra Voss @csvoss
10K Followers 1K Following engineeress ✨ Member of Technical Staff @openai serious play // notice your curiosity
Cathy Wu @wucathy
2K Followers 872 Following Accepting PhD students and postdocs. Faculty at @MIT. Intelligent Multi-agent Coordination. ML. Autonomy. Mobility. Reinforcement Learning. Public Policy.
Ruiqi Zhong @ZhongRuiqi
2K Followers 698 Following 5th Year Ph.D. @BerkeleyNLP, Columbia'19. part time working for @AnthropicAI . Supervising machines to do what I can't do.
Tatsunori Hashimoto @tatsu_hashimoto
6K Followers 202 Following Assistant Prof at Stanford CS, member of @stanfordnlp and statsml groups; Formerly at Microsoft / postdoc at Stanford CS / Stats.
John Cherian @jjcherian
2K Followers 353 Following Statistics PhD student @Stanford and @HertzFoundation fellow | Quantifying uncertainty @WashingtonPost | Formerly at @DEShawResearch
Andrew Ilyas @andrew_ilyas
2K Followers 167 Following Machine Learning PhD student at MIT, advised by Aleksander Madry and Costis Daskalakis.
Alex Tamkin 🦣 @AlexTamkin
4K Followers 1K Following machine learning, science & society @AnthropicAI | prev: phd @StanfordAILab, @stanfordnlp
Jonathan Huggins @jhhhuggins
968 Followers 282 Following Assistant Professor of Statistics at Boston University | Co-founder @macovidvaxhelp | dad, rower, he/him | @StatisticsBU | @BU_CDS | @BU_Computing | @BU_Tweets
Wei Hu @weihu_
2K Followers 1K Following Assistant professor @UMich; previously @UCBerkeley @Princeton @Tsinghua_Uni. Building the theoretical and scientific foundations of deep learning.
Yasaman Bahri @yasamanbb
5K Followers 938 Following Research Scientist @GoogleDeepMind // ML + physics + quantum materials // Ph.D. theoretical cond matt physics @UCBerkeley.
Mihaela Curmei 🇺�.. @mihaela_curmei
141 Followers 676 Following EECS PhD student @Berkeley previously @Microsoft and @Princeton
Michi Yasunaga @michiyasunaga
3K Followers 867 Following CS PhD @Stanford working on language models and multimodal models. Previously @Meta @GoogleDeepMind @Yale
Helen Toner @hlntnr
21K Followers 1K Following Interests: China+ML, natsec+tech, brains+words+absurdity | Current: @CSETGeorgetown (opinions my own) | Former: @open_philIn a new preprint with Jarek Blasiok, @rares_buhai, David Steurer, we show a surprisingly simple greedy algorithm that can list decode planted cliques in the semirandom model at k~sqrt n log^2 n --essentially optimal up to log^2 n. This ~resolves @JacobSteinhardt's open question.
I've some exciting news! Like much of my recent work, also inspired by Uri Feige's conjectures. With Rares Buhai & David Steurer (during a beautiful Zurich summer), we found new algos for semirandom planted clique at thresholds approaching "usual" PC. arxiv.org/abs/2212.05619
🌟Easy-to-Hard Generalization: Scalable Alignment Beyond Human Supervision 🌟 arxiv.org/abs/2403.09472 How can we keep improving AI systems when their capabilities surpass those of human supervisors? (1/n)
oh no we need to RLHF the protein LLMs to make sure they aren't specieist please choose which protein sequence is less speciest so we can align AI with human & non-human values sequence 1: MRWQEMGYIF YPRKLR sequence 2: LPDCKVMVHD PHSLA
Protein language models (pLMs) can give protein sequences likelihood scores, which are commonly used as a proxy for fitness in protein engineering. But what do likelihoods encode? In a new paper (w/ @JacobSteinhardt) we find that pLM likelihoods have a strong species bias! 1/
[1/5] Introducing VisDiff - an #AI tool that describes differences in image sets with natural language. VisDiff can summarize model failures, compare models, find nuanced dataset differences, discover what makes an image memorable, and so much more! …derstanding-visual-datasets.github.io/VisDiff-websit…
[5/5] This work wouldn't be possible without the amazing collaboration between @Zhang_Yu_hui @wxh1996111 @syeung10 at @StanfordAILab and @ZhongRuiqi @trevordarrell @JacobSteinhardt @profjoey at @berkeley_ai. Paper: arxiv.org/abs/2312.02974 Code: github.com/Understanding-…
@JacobSteinhardt @DhruvMadeka In particular, I hope that people summarizing/glossing over your article don't get that wrong.
@JacobSteinhardt @DhruvMadeka I like the overall analysis. I think that the move of noticing that AIs might share some characteristics with pandemics, in that AIs might be self-replicating, is an inside-view move, and I don't feel great about characterizing that as a reference class analysis.
@haldaume3 General papers on AI safety leave me wanting more detail and less speculation. I suggest looking at the technical papers people publish. Papers by @DanHendrycks and @JacobSteinhardt would be one place to look.
@vrtejus You might also like @JacobSteinhardt's essay, "Research as a Stochastic Decision Process", which goes into much more depth with productivity tips! 😁 docs.google.com/document/d/1KC…
The married mathematicians Eric Larson and Isabel Vogt often found themselves discussing ideas after dinner, working through problems on the chalkboards they have in their home. The pair recently proved a centuries-old question about algebraic curves. quantamagazine.org/old-problem-ab…
Highly recommend @JacobSteinhardt 's trilogy on AI forecasting. * bounded-regret.ghost.io/ai-forecasting/ * bounded-regret.ghost.io/ai-forecasting… * bounded-regret.ghost.io/scoring-ml-for… It's calming to see the process used to cut through some of the chaos, as well how uncertain even expert distributions are.
I think the new Anthropic paper is cool. It has me updating positively toward this. But here are some things I would have liked to see. - Not measuring success by simply having a human look at results and judge how "interpretable" they are. This type of evaluation methodology…
I was excited to see our paper on challenges with RLHF quoted in this Globe and Mail article. The quote was attributed to me since I was the first listed author. But it was from the part written by @FreedmanRach. theglobeandmail.com/business/artic…
New paper!! We found a pattern in how NNs extrapolate: as inputs become more OOD, model outputs tend to go towards some “average”-like prediction. What is this “average”-like prediction? Why does this happen? Can we leverage this to better handle OOD inputs? (Spoiler: Yes!) 🧵:
New J. Steinhardt blogpost just dropped. I highly recommend. Jacob has a very careful way of thinking about messy topics that I personally really like. I particularly recommend his posts "More is Different" and "Complex Systems are Hard to Control".
Over the past two years, I and many other forecasters registered predictions about the state-of-the-art accuracy on ML benchmarks in 2022-2025. In this blog post, I evaluate the predictions for 2023: bounded-regret.ghost.io/scoring-ml-for…
Check the rest of the program (recorded) by @chrmanning @ilyasut @Diyi_Yang @colinraffel @boknilev @YejinChoinka @prfsanjeevarora @MilesCranmer @JacobSteinhardt @spiantado @JitendraMalikCV @SuryaGanguli @lschmidt3 & many others simons.berkeley.edu/workshops/larg… x.com/simonsinstitut…
How can we audit language models for rare behaviors that typical evaluation might miss (e.g., find prompts that generate “Elizabeth Warren” or French words that generate English words)? In our #ICML23 paper, we aim to uncover these behaviors via discrete optimization 🧵
[1/9] Large Language Models (LLMs) can mimic humans to explain human decisions. But can they explain THEMSELVEs? How to evaluate explanations along this axis? Check out our work “Do Models Explain Themselves? Counterfactual Simulatability of Natural Language Explanations”!
Trends for United States
You might like
