Adrià Garriga-Alonso (hiring @ far.ai) @AdriGarriga
Research Scientist at FAR AI (@farairesearch), towards AI beneficial to everyone. agarri.ga Berkeley, California Joined February 2014-
Tweets748
-
Followers649
-
Following576
-
Likes5K
Do models need to reason in words to benefit from chain-of-thought tokens? In our experiments, the answer is no! Models can perform on par with CoT using repeated '...' filler tokens. This raises alignment concerns: Using filler, LMs can do hidden reasoning not visible in CoT🧵
Good on @GoogleDeepMind for following through on these commitments. Would like to see an explanation from @OpenAI & @AnthropicAI for apparent breach of this commitment.
Good on @GoogleDeepMind for following through on these commitments. Would like to see an explanation from @OpenAI & @AnthropicAI for apparent breach of this commitment.
We modeled AI learning from (un)reliable human teachers. But what happens when humans disagree about what the AI should do altogether? In a new position paper, we propose addressing conflicting preferences using social choice theory. Out now on arxiv! arxiv.org/abs/2404.10271
We modeled AI learning from (un)reliable human teachers. But what happens when humans disagree about what the AI should do altogether? In a new position paper, we propose addressing conflicting preferences using social choice theory. Out now on arxiv! arxiv.org/abs/2404.10271
.@TheZvi woke up and chose violence: "Sam Altman is not playing around. He wants to build new chip factories in the decidedly unsafe and unfriendly UAE. He wants to build up the world’s supply of energy so we can run those chips. What does he say these projects will cost? Oh,…
.@TheZvi woke up and chose violence: "Sam Altman is not playing around. He wants to build new chip factories in the decidedly unsafe and unfriendly UAE. He wants to build up the world’s supply of energy so we can run those chips. What does he say these projects will cost? Oh,… https://t.co/77BDxK3uLg
Sam Altman before founding OpenAI
Awesome way to get incentivize good epistemic uncertainty in models during training!
New paper: How can you tell when a model is hallucinating? Let it cheat! An expert doesn't need to cheat, so if your model learns to cheat, there must be something it doesn't know. Our general new approach for measuring uncertainty: arxiv.org/abs/2402.08733
If you want to know about PauseAI US's doings, follow the twitter account!
🎉 They're live! Dive into #AIAlignment at the #AlignmentWorkshop with videos now on YouTube & our site, all with captions & transcripts. 📺 For more insights, check out our blog post. ✨Links below 🔗👇Be inspired, engage, and share your favorite insights!
According to AI Impact's survey, scientists' estimated time to AI milestones has gone down a lot on average. Tracks with my experience!
According to AI Impact's survey, scientists' estimated time to AI milestones has gone down a lot on average. Tracks with my experience!
@cameron_pfiffer You are correct that SF just stopped enforcing traffic laws
“Let’s focus on today’s problems, not hypothetical future ones” is the worst counter to existential risk arguments. You could analogously argue against climate change mitigation and a host of other future-oriented concerns. Let’s actually assess the likelihood of AI apocalypse.
🎉 Reflecting on a fantastic #NeurIPS2023 #AIAlignment Workshop! 🚀 🙌 149 attendees energized the main event 🌃 500+ at our Monday social 🧠 12 talks, 25 lightning talks 🔑 Keynote by Yoshua Bengio 🤔 What inspired you the most? Share your thoughts!
If you felt disturbed by the OpenAI governance debacle, and you work in AI, you might be tempted to work on "alignment" to help reduce your worries that AI will get out of control. But why not channel your technical abilities to work directly on something that helps with…
Which is why at @farairesearch we're running a project on empirically testing whether language models have goals. (my guess is right now they don't but it'll change)
Which is why at @farairesearch we're running a project on empirically testing whether language models have goals. (my guess is right now they don't but it'll change)
Join us for the alignment social at #NeurIPS2023!
Join us for the alignment social at #NeurIPS2023!
I have gotten more requests than I’d expect for introductions to my contract artist, DALL-E.
Today in Bits about Money: an in-depth explanation of the shorthand that I've used for a few years about crypto jurisdictional gamesmanship. Binance and CZ, major practitioners of the Bond villain compliance strategy, are having a bit of a rough week. bitsaboutmoney.com/archive/bond-v…
4. Making memes come to life using the new Stable Diffusion Video
🧵Announcing GPQA, a graduate-level “Google-proof” Q&A benchmark designed for scalable oversight! w/ @_julianmichael_, @sleepinyourhat GPQA is a dataset of *really hard* questions that PhDs with full access to Google can’t answer. Paper: arxiv.org/abs/2311.12022
Richard Ngo @RichardMCNgo
35K Followers 1K Following What would we need to understand in order to design an amazing future? Figuring that out @openaiVidhi Lalchand @VRLalchand
1K Followers 1K Following Eric and Wendy Schmidt Center Postdoctoral fellow @Schmidt_Center @broadinstitute @MIT PhD @CambridgeMLG @Cambridge_Uni 🇺🇲 via 🇬🇧 & 🇮🇳David Krueger @DavidSKrueger
13K Followers 4K Following Cambridge faculty - AI alignment, deep learning, and existential safety. Formerly Mila, FHI, DeepMind, ElementAI, AISI.Sam Power @sp_monte_carlo
17K Followers 7K Following Lecturer in Maths & Stats at Bristol. Interested in probabilistic + numerical computation, statistical modelling + inference. (he / him)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)Andreas Kirsch 🇮�.. @BlackHC
9K Followers 5K Following Past: 🧑🎓 DPhil @AIMS_oxford @ExeterCollegeOx @UniofOxford (4.5yr) 🧙♂️ RE @DeepMind (1yr) 📺 SWE @Google (3yrs) 🎓 @TU_Muenchen 👤 Fellow @nwspkJames Hensman @jameshensman
7K Followers 2K Following Machine learner. Building big Bayesian models @microsoft. Views my own. he/him.Sebastian Ober @sebastian_ober
501 Followers 391 Following Senior Scientist for ML in Biologics Engineering at @AstraZeneca. Previously PhD student and @Gates_Cambridge scholar at @CambridgeMLG. Views are my own.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.Javier Antorán @JaviAC7
769 Followers 439 Following Interested in Bayesian Inference and Molecular Dynamics @CambridgeMLG.Michael Hutchinson (@.. @MHutchinson141
687 Followers 335 Following PhD student @OxfordStats / @OXCSML supervised by @yeewhye and @wellingmax. Probabilistic ML, geometric ML and their interestion. Interned @DeepMind @Qualcommgavin leech @g_leech_
4K Followers 421 Following the subject of criticism @ArbResearch, @Bristol_AI_CDT, ESPRMarius Hobbhahn @MariusHobbhahn
2K Followers 996 Following Director/CEO at Apollo Research @apolloaisafety Ph.D. student of Machine Learning @PhilippHennig5; AI safety/alignmentJan Leike @janleike
44K Followers 322 Following ML Researcher, co-leading Superalignment @OpenAI. Optimizing for a post-AGI future where humanity flourishes.James Allingham @JamesAllingham
1K Followers 459 Following RS @GoogleDeepMind | Machine Learning PhD @CambridgeMLG | 🇿🇦Alexander Terenin @avt_im
6K Followers 950 Following Machine learning, artificial intelligence, decision theory | anti-ideological | thinking carefully about incentives | Assistant Research Professor @CornellMiri Zilka @MiriZilka
291 Followers 414 Following @LeverhulmeTrust Research Fellow in Trustworthy Machine Learning at @CambridgeMLG. CRA @Kings_College and Associate Fellow @LeverhulmeCFI.Jaime Sevilla @Jsevillamol
2K Followers 322 Following Director of @EpochAIResearch. Technological forecasting and trends in Machine Learning.MabelTom @w231m7ca936FDZM
0 Followers 205 FollowingSébastien Darses @DarsesSebastien
3K Followers 3K Following Assoc. Prof. (he/him 🏳️🌈), Visiting @CRM_Montreal @CNRS_INSMI @CNRS 📚 Probability, Statistics, Analytic Number Theory 👨🎓Teaching project @highkholleTessa Bohmann @TesBohma
84 Followers 5K FollowingDaniel Johnson (@ ICL.. @_ddjohnson
2K Followers 579 Following Researcher at @GoogleDeepMind. PhD student at @VectorInst / @UofT. Building tools to study neural nets and find out what they know. He/him.Frank Yan @yantao
140 Followers 1K Following Full Stack Software Engineer specializing in Web and Blockchain Technologies.Arif Ahmad @arif_ahmad_py
309 Followers 7K Following All things AI, Computer Science and Circuits! Prev. @GoogleAISri Mahaguhan @SriMahaguhan
32 Followers 190 FollowingRyan Kidd @ryan_kidd44
971 Followers 848 Following Co-Director, @MATSprogram + Co-Founder, https://t.co/26oYPZwxVx | PhD in physics | Accelerate AI alignment + build a better future for allDianaBauer @bZN3dWVaQ9IauLd
0 Followers 195 FollowingKrueger AI Safety Lab @kasl_ai
276 Followers 67 Following We are a research group at the University of Cambridge focused on avoiding catastrophic risks from AI.SSF @SureshShuklaFan
3K Followers 4K Following Parody Account || Welcome to the Suresh Shukla FanClub Stay tuned for regular updates. DM o contacto para promociones y empleos. Ciudad de México y Montréalbczcu4ttj7zxa4j8 @3ufk5euroy
5 Followers 1K Following We first transfer USDT to you TRC20, you return 90% to BEP20, you get 10% , 2K per day Our co hv a large amt of USDT need to from TRC20 convert to BEP20 networkBen Lerner @benjamin_lerner
173 Followers 276 Following Ads ML @DoorDash, prev. @twosigma, @snap | studying AI safety @BlueDotImpact | @USAPowerlifting competitorGabriele Sarti @gsarti_
2K Followers 2K Following PhD Student @GroNLP 🐮, core dev of @InseqLib (https://t.co/tTjrg26ygQ). Interpretability ∩ HCI ∩ #NLProc. Prev: @AmazonScience, @Aindo_AI, @ItaliaNLP_Lab.Anna ⛓️🫧🪽 @annakayshive
110 Followers 341 Following Elle Woods of longtermism 💅 Vapra clan gelfling gf 📖 Aspiring trad ⛪ Aspiring MILF 🍼 cyborg designer 🤖🎨Chris Cundy @ChrisCundy
1K Followers 194 Following PhD student at Stanford AI Lab, supervised by Stefano Ermon. Hopefully making AI benefit humanity. Anonymous feedback: https://t.co/Wh3rHMsRnmRocket Drew @rocketalignment
58 Followers 127 Following AI Journalism @Tarbell_Fellows, Community Manager @MATSProgramMarco Molinari @marco__molinari
45 Followers 322 Following Founder / Lead of https://t.co/LaEGUjKBpa | Student @LSEDataScience 💂 | Ex. Machine Learning Research FellowAdeyemiSolomon Adegbo.. @AdeyemisolomonA
139 Followers 2K Following Web maestro 🚀 | crafting digital experiences✨ | Transforming visions into sleek websites 💡 #WebDesign #UXUI #DigitalExperience #ResponsiveDesign #CreativeMindbjolo @bjolo8442
120 Followers 867 Following沈东 @579ls
0 Followers 76 FollowingCoding But Still Aliv.. @CbsaSciencehub
25 Followers 518 Following Coding But Still Alive - that’s my passion. I am a Data Scientist & ML Engineer with a special interest in advanced AI and Deep Learning. PhD in Bioinformatics.Juan Hmmm @JuanAH03488233
76 Followers 3K FollowingClaire Short @rocksandbugs
3 Followers 48 FollowingRogan Inglis @RoganInglis
39 Followers 569 Following Co-Founder / Senior Machine Learning Engineer at IntelistyleAlgeia @Algeia17812
18 Followers 70 FollowingDanny Halawi @dannyhalawi15
170 Followers 290 Following masters student at @berkeley_ai advised by @JacobSteinhardt. Interested in interpretability, scalable oversight, and forecasting.Vaibhav Raj @vrcoder045
38 Followers 1K Following Comp. Sci. Senior at IIT Bombay, upcoming SWE, ML enthusiastEvan Anders @evanhanders
82 Followers 140 Following AI Safety / Mech Interp postdoctoral scholar @KITPUCSB. Former astrophysical fluid dynamicist @Northwestern (CIERA) and @CUBoulder.pawann k. @pawaniiit
223 Followers 4K Following Prof., PhD, Inria, France, Postdoc KU Leuven, Fraunhofer ITWM, FU Berlin. I like Machine learning and mathematics.Nikhil Reddy @nikhil_reddy_cs
261 Followers 5K Following Ph.D student at the University of Queensland IIT Delhi Research Academy | Research areas: Data science, Computer visionFrancesc Lluis @francesclluis_
228 Followers 493 Following Deep learning for audio signal processing and acoustics @BangOlufsen.Guillem Cucurull @g_cucurull
402 Followers 496 Following Machine Learning and Computer Vision, doing cool things at @paperswithcode (@MetaAI)Fabian Falck @fabianfalck
218 Followers 894 Following PhD student in ML @UniofOxford @oxcsml @OrielOxford Prev. @MSFTResearch @AmazonScience @imperialcollege @KITKarlsruhe (Probabilistic) Generative ModelsChrisy Bornberg @variint
2K Followers 2K Following Interpretability of modular networks for retinal disease understanding: 👁 @snec_seri @NTUsg | 👁 @MedUni_Wien | 🤰🏻 @WEISS_UCL | she/herOliver Daniels-Koch @Oliver_ADK
60 Followers 293 FollowingBurny — Effective O.. @burny_tech
14K Followers 6K Following Transhuman engineer in singularity! Lover of AI & omnidisciplionary metamathemagics! Shapeshifting metafluid! Hypercuriousia! Omniperspectivity! Freedom 4 all!Arushi GK Majha @arushimajha
247 Followers 921 Following scientist 💭💻🧠 ai/ml+medchem phd @Cambridge_Uni • director @WiMLworkshop • she/godless/non-mum 🧚♀️ • uc/oc/dvij/brit 💩• sentient welfare, near+longterm 🖤Alan Chan @_achan96_
858 Followers 1K Following PhD student @Mila_quebec || Research Scholar @GovAI_ || AI safety || 🇨🇦Yann LeCun @ylecun
713K Followers 718 Following Professor at NYU. Chief AI Scientist at Meta. Researcher in AI, Machine Learning, Robotics, etc. ACM Turing Award Laureate.Richard Ngo @RichardMCNgo
35K Followers 1K Following What would we need to understand in order to design an amazing future? Figuring that out @openaiNeurIPS 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].Andrej Karpathy @karpathy
981K Followers 905 Following 🧑🍳. Previously Director of AI @ Tesla, founding team @ OpenAI, CS231n/PhD @ Stanford. I like to train large deep neural nets 🧠🤖💥Google DeepMind @GoogleDeepMind
945K Followers 275 Following We’re a team of scientists, engineers, ethicists and more, committed to solving intelligence, to advance science and benefit humanity.Clément Canonne @ccanonne_
31K Followers 928 Following Senior Lecturer @Sydney_Uni. Postdocs @IBMResearch, @Stanford; PhD @Columbia. Converts ☕ into puns: sometimes theorems. He/him. @[email protected]Vidhi Lalchand @VRLalchand
1K Followers 1K Following Eric and Wendy Schmidt Center Postdoctoral fellow @Schmidt_Center @broadinstitute @MIT PhD @CambridgeMLG @Cambridge_Uni 🇺🇲 via 🇬🇧 & 🇮🇳Michael A Osborne @maosbot
33K Followers 1K Following Dad, spouse, Professor of Machine Learning @UniofOxford, Co-Founder Mind Foundry, Director @aims_oxford. Bayes, Long Covid, porridge, AI must be good for humansAnthropic @AnthropicAI
264K Followers 26 Following We're an AI safety and research company that builds reliable, interpretable, and steerable AI systems. Talk to our AI assistant Claude at https://t.co/aRbQ97uk4d.François Chollet @fchollet
470K Followers 770 Following Deep learning @google. Creator of Keras. Author of 'Deep Learning with Python'. Opinions are my own.David Krueger @DavidSKrueger
13K Followers 4K Following Cambridge faculty - AI alignment, deep learning, and existential safety. Formerly Mila, FHI, DeepMind, ElementAI, AISI.Sam Power @sp_monte_carlo
17K Followers 7K Following Lecturer in Maths & Stats at Bristol. Interested in probabilistic + numerical computation, statistical modelling + inference. (he / him)Eric Jang @ericjang11
69K Followers 3K Following physical AGI at 1X. Author of "AI is Good for You" https://t.co/eFg4WXhg0pJose Miguel Hernánde.. @jmhernandez233
4K Followers 120 Following Professor of Machine Learning, University of Cambridge, UK.Ferenc Huszár @fhuszar
40K Followers 1K Following Secular Bayesian. Associate Professor in Machine Learning @Cambridge_CL. Talent aficionado at https://t.co/RbJkoLguey Alum of @Twitter, Magic Pony and @BaldertonAndreas Kirsch 🇮�.. @BlackHC
9K Followers 5K Following Past: 🧑🎓 DPhil @AIMS_oxford @ExeterCollegeOx @UniofOxford (4.5yr) 🧙♂️ RE @DeepMind (1yr) 📺 SWE @Google (3yrs) 🎓 @TU_Muenchen 👤 Fellow @nwspkEliezer Yudkowsky ⏹.. @ESYudkowsky
175K Followers 89 Following The original AI alignment person. Missing punctuation at the end of a sentence means it's humor. If you're not sure, it's also very likely humor.James Hensman @jameshensman
7K Followers 2K Following Machine learner. Building big Bayesian models @microsoft. Views my own. he/him.Anna ⛓️🫧🪽 @annakayshive
110 Followers 341 Following Elle Woods of longtermism 💅 Vapra clan gelfling gf 📖 Aspiring trad ⛪ Aspiring MILF 🍼 cyborg designer 🤖🎨Sebastian Ober @sebastian_ober
501 Followers 391 Following Senior Scientist for ML in Biologics Engineering at @AstraZeneca. Previously PhD student and @Gates_Cambridge scholar at @CambridgeMLG. Views are my own.Chris Cundy @ChrisCundy
1K Followers 194 Following PhD student at Stanford AI Lab, supervised by Stefano Ermon. Hopefully making AI benefit humanity. Anonymous feedback: https://t.co/Wh3rHMsRnmYawen Duan @yawen_duan
285 Followers 413 Following Concordia AI https://t.co/Pe2BhjbbE0 | AI Safety & Governance | ML MPhil @Cambridge_Eng @kasl_ai | ex-intern @CHAI_BerkeleyAccepted papers at TM.. @TmlrPub
3K Followers 2 FollowingJan Brauner @JanMBrauner
776 Followers 313 Following PhD student in ML. University of Oxford, @OATML_Oxford.r @theorizur
636 Followers 571 Following straight lines gods' worshiper · human disempowerment is natural selection's default outcomeAleksander 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.Ryan Kidd @ryan_kidd44
971 Followers 848 Following Co-Director, @MATSprogram + Co-Founder, https://t.co/26oYPZwxVx | PhD in physics | Accelerate AI alignment + build a better future for allXin Cynthia Chen @XinCynthiaChen
351 Followers 337 Following Direct PhD student @ETH_en, with research focus on AI Safety and Alignment. Formerly at @CHAI_Berkeley.Bilal Chughtai 🇵�.. @bilalchughtai_
591 Followers 583 Following ai safety | mechanistic interpretability | cambridge mmathFAR AI @farairesearch
1K Followers 19 Following Ensuring AI systems are trustworthy and beneficial to society by incubating new AI safety research agendas.Karla Ortiz @kortizart
90K Followers 6K Following Karla is a Puerto Rican artist who works on Films (MCU, ILM,HBO), Games, TV, Covers, Fine art, etc. Passionate advocate for better artist industries+ rights ✌️kipply @kipperrii
8K Followers 826 Following "drop the forest nymph act we know how much gdp you generate" - @mnovendstern | alt @kipperriiiiSiméon @Simeon_Cps
7K Followers 1K Following Creating more common knowledge on AI risks, one tweet at a time. Founder in Paris. AI auditing, standardization & governance.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.GoldieSilverman @GoldieSilverman
10K Followers 40 FollowingTarek Mansour @mansourtarek_
33K Followers 2K Following ceo @Kalshi. ex MIT, Citadel, Palantir. I like markets. https://t.co/lwkzyUqeAxNino Scherrer @ninoscherrer
596 Followers 2K Following Interested in rigorous LLM evals, synthetic data, causality, robustness & AI safety/ethics | RS @PatronusAI | Prev: @VectorInst, @Mila_Quebec, @MPI_IS, @ETH_enFrances Lorenz @frances__lorenz
4K Followers 538 Following ✨ I share my feelings, post lil jokes for the girlies, and often discuss effective altruism ✨ I also work on the EA Global team at CEA (views my own)Michał Zając @Michal_Zajac_
127 Followers 228 Following Interested in AI safety. Research Engineer at FAR AI, PhD student at Jagiellonian University, previously intern at Google Brain, UofT, KU LeuvenNitarshan Rajkumar @nitarshan
813 Followers 1K Following Adviser to the Secretary of State @scitechgovuk. Cofounder @aisafetyinst. Co-created AI Safety Summit and UK AI Research Resource. PhD @cambridge_cldepths of wikipedia! @depthsofwiki
889K Followers 4K Following Hello I am @anniierau Please take away my blue check! I did not ask for it!Metal Català @metalcatala
1K Followers 1K Following Aquí trobaràs tota la informació sobre el metal en català 🔥#metalcatalà #metalencatalà🔥Holly ⏸️ Elmore @ilex_ulmus
4K Followers 459 Following Dedicated to the protection and thriving of sentient beings. PhD in evo bio. Executive Director of @PauseAIUS. Opinions not necessarily those of the org.Cas (Stephen Casper) @StephenLCasper
3K Followers 1K Following #AI safety & responsibility. PhD Candidate @ #MIT_CSAIL.Tom Lieberum @lieberum_t
949 Followers 178 Following Trying to reduce AGI x-risk by understanding NNs Interpretability RE @DeepMind BSc Physics from @RWTH GWWC pledgee @ https://t.co/Vh2bvwhuwdEvan Hubinger @EvanHub
4K Followers 1K Following Alignment stress-testing team lead @AnthropicAI. Opinions my own. Previously: MIRI, OpenAI, Google, Yelp, Ripple. (he/him/his)Pete Mandik @petemandik
6K Followers 1K Following Freestanding utility protein. https://t.co/apJUV1Do1JManifold @ManifoldMarkets
8K Followers 296 Following The largest prediction market platform. Bet on politics, tech, sports, and more. Create your own play-money market. Not crypto.Ajeya Cotra @ajeya_cotra
6K Followers 286 Following AI could get really powerful soon and I worry we're underprepared. Analysis+grantmaking in AI alignment @open_phil (views my own), editor+writer @plannedobs.Jacob Steinhardt @JacobSteinhardt
7K Followers 67 Following Assistant Professor of Statistics, UC BerkeleyMiranda Zhang @mirandahzhang
1K Followers 1K Following suffering reduction, AI safety, animal welfare, affordable housing. 💖 opinions my own.Marius Hobbhahn @MariusHobbhahn
2K Followers 996 Following Director/CEO at Apollo Research @apolloaisafety Ph.D. student of Machine Learning @PhilippHennig5; AI safety/alignmentDan Hendrycks @DanHendrycks
17K Followers 81 Following • Director of the Center for AI Safety (https://t.co/ahs3LYCpqv) • GELU/MMLU/MATH • PhD in AI from UC Berkeley https://t.co/rgXHAnYAsQ https://t.co/nPSyQMaY9bAlex Turner @Turn_Trout
1K Followers 39 Following Research scientist on the scalable alignment team at Google DeepMind. All views are my own.Yoshiaki Araki 荒木.. @alytile
3K Followers 3K Following 日本テセレーションデザイン協会 代表 ミラクル エッシャー展 スーパーバイザー/ニュートン別冊 図形編 エッシャー記事監修/ 映画「エッシャー 視覚の魔術師」広報翻訳協力/ マーブルシュッド 「Tessellation」監修 / 著書 「M.C.エッシャーと楽しむ算数・数学パズル」Vaidehi is in NYC! @vaidehiagrwalla
908 Followers 516 Following Product @ Momentum (+ https://t.co/lWG5xZhohI🍍). I like updating (my beliefs). 🇸🇬🇮🇳Haoxing Du @haoxingdu
88 Followers 125 Following Only ever worked on applied linear algebra. Effective Altruist. Dazed and confused, but trying to continue.Yo Shavit @yonashav
4K Followers 831 Following policy for v smart things @openai. Past: CS PhD @HarvardSEAS/@SchmidtFutures/@MIT_CSAIL. Tweets my own; on my head be it.Haydn Belfield @HaydnBelfield
4K Followers 2K Following @Cambridge_Uni researcher. Tweets about international security, AI governance, pandemics, nukes and climate change. @CSERCambridge & @LeverhulmeCFI@lucyfarnik I think publicly leaving and saying he they are not competent or trustworthy (not to mention forcing them to replace him) is way more impactful than moving within whatever narrow latitude he had at the org
😂😂 I’m reminded of a course offered called “Real simple groups”. 100 people showed up to the first lecture (presumably expecting it to be really simple, nevermind the grammatical error). Only five returned for lecture 2.
A math department at a major university in the US is now offering a semester long course on “game theory”. Yes, you read that right. Games. The things tiny children play. The dumbing down of America continues…
An update on our work on SAEs. Stay tuned for our upcoming SAE Pareto improvement too… :)
Announcing a progress update from the @GoogleDeepMind mech interp team! Inspired by @AnthropicAI's excellent monthly updates, we share a range of updates on our work on Sparse Autoencoders, from signs of life on interpreting steering vectors with SAEs to improving ghost grads.
Someone complimented me for a random modal logic lecture I recorded in 2017. It made my day!
We modeled AI learning from (un)reliable human teachers. But what happens when humans disagree about what the AI should do altogether? In a new position paper, we propose addressing conflicting preferences using social choice theory. Out now on arxiv! arxiv.org/abs/2404.10271
RLHF typically assumes that all training feedback comes from a single teacher, but teachers can disagree up to 37% of the time in practice. In our new paper, we introduce active teacher selection to learn from different teachers. (1/n)
viridis
Found a baby fox on my dog walk, crying and walking up to people for help. Had no choice but to take it home and it immediately settled into my dog’s crate. The fox rescuers are on their way now my good gosh look at that face
it is the space year 2024: The Kuomintang are pro-communist. The Republicans support Russia. The feminists are biodeterminists. Teenagers do not party.
@jessesingal It is at least sociologically very interesting that this is one of the things that feminists who are now called “gender-critical” spent decades denying - and now it’s strongly associated with their side of the debate! Funny old world.
cute (if a bit under-explained): (from arxiv.org/abs/2311.11924)
My effective altruism cause area this year is weapons for Ukraine. Most effective way to improve long term walys in Eastern Europe for sure 🔥💪
I'd like to sponsor some weapons for Ukraine. Where do I donate? Not for humanitarian aid, but for the other kind. Please DM.
Briefly, I want to address the issue of who is to blame. Easy — the people behind the attack. Lasse, the maintainer of xz, was the target of a patient intelligence campaign that invested more resources into subverting him than anyone invested into his project.
Australians take Easter weekend seriously
Does anyone want to defend Zack Robinson's recent article in the WaPo to me? There is an EA Forum thread about it, but I am just really shocked how much it really just seems like a vacuous puff-piece, and I am pretty confused with the WaPo published it, and would like to see…
OpenAI: “The mission of OpenAI is to ensure artificial general intelligence benefits all of humanity.” This isn’t very consistent with that. 🤔 Meanwhile OpenAI engineers often make >800k a year.
For years I’ve been interviewing data annotation workers who are the lifeblood of the AI industry. For years I’ve heard the same story: the platforms they work for wield total power, leaving them precarious & vulnerable to exploitation. A horrible example of this just happened 1/
...multiple ICML submissions mentioning in passing how you can use chain-of-thought to figure out why a model did what it did, without any awareness that it's not necessarily faithful.
I'm excited to release Prisma, a mechanistic interpretability library for multimodal models like CLIP and ViTs. Incubated at @tyrell_turing's lab & in collab with @NeelNanda5. Recent mech interp work has focused on language, but many techniques transfer. Behold, the dogit lens:
@ohabryka @AnthropicAI It's plausible to me that it's just unknown whether Claude 3 or GPT-4 are SOTA in a given task. We'll probably know more in a few months
Yesterday’s “too dangerous to release” is today’s 98% off. There’s no moat, and the price wars are on.
Did anthropic just kill every small model? If I'm reading this right, Haiku benchmarks almost as good as GPT4, but its priced at $0.25/m tokens It absolutely blows 3.5 + OSS out of the water For reference gpt4 turbo is 10m/1m tokens, so haiku is 40X cheaper.