Eric Ho @eric_ho
Co-Founder / CEO @GoodfireAI - AI interpretability research company goodfire.ai San Francisco Joined September 2011-
Tweets396
-
Followers3K
-
Following404
-
Likes3K
it's really good that openAI and huggingface were so transparent and collaborative on this safety issue, i think these types of problems are going to be increasingly challenging for any one organization to solve
this is pretty nuts. i think this is the tip of the iceberg for needing to focus on agentic AI safety. models will only get more capable from here, and RL incentivizes getting the right answer at any cost this is a large reason why we're obsessed with 'intentional design', in other words, making sure models are trained to do the right things for the right reasons
we had a significant security incident during evaluation of our models. we are sharing what we have learned so far. thanks to @huggingface for the partnership on this. openai.com/index/hugging-…
we're hiring a lead for model training & training research. if you want to make models great at interpretability, and push the frontiers of intentional design, apply here: job-boards.greenhouse.io/goodfire/jobs/…
Can we finally see inside AI's black box? 👀 @eric_ho, CEO of @GoodfireAI, explains how AI interpretability is helping researchers understand what happens inside neural networks. 📺 YouTube: link.theneurondaily.com/r/9mHDDm 🎧 Spotify: link.theneurondaily.com/r/cPbcJb 🎙️ Apple Podcast: link.theneurondaily.com/r/ak0wxY #AIInterpretability #NeuralGeometry #MachineLearning #Goodfire #TheNeuron
@theneurondaily @GoodfireAI thanks for having me! fun conversation
glad to see a strong open weight model, also excited to inspect the internal representations across modalities
Today, we are introducing Inkling. Inkling reasons efficiently across text, image, and audio modalities. We are making the full weights available. thinkingmachines.ai/news/introduci… Available today for fine-tuning on Tinker. Play with it in the Inkling Playground. 🧵
Almost all my work is done via Silico now :D
> replicate J-space on GLM 5.2 > train a reward model and run RL to reduce hallucinations > show me how this model makes cancer predictions Using our platform Silico is like having a team of AI researchers ready to run experiments like these. Private beta is open now. 🧵 (1/6)
i am consistently amazed by the quality of the experiments that silico can run. excited to kick this off in private beta! we're also doing grants to select AI safety and academic researchers. DM me if you're working on something interesting
> replicate J-space on GLM 5.2 > train a reward model and run RL to reduce hallucinations > show me how this model makes cancer predictions Using our platform Silico is like having a team of AI researchers ready to run experiments like these. Private beta is open now. 🧵 (1/6)
i think this is a good idea!
just welcomed chris earls, cornell engineering professor, to the team! i'm particularly excited about his work to unlock scientific creativity in frontier AI with interpretability we've hired several professors at @GoodfireAI because we're investing heavily in foundational research to discover the science of neural networks. if this is work that you're interested in, join us!
Goodfire's Silico decided to show me this result of how representation of days of the week emerges over pretraining
i think it’s important to be default loving rather than default fearful of AI. i feel the same way
I know it might kill everyone, but I can't help loving AI. All my life I wanted to talk to an alien intelligence, but I never thought I'd really get the chance to.
we can use interp to see what models truly believe versus what they say. often a surprising difference
Can LLMs predict the next World Cup champion? Goodfire partnered with @EternisAI to improve how LLM forecasters use available evidence and manage uncertainty. We found models were overconfident in their predictions – but probes significantly improved calibration. (1/6)
Never been more bullish on design and Figma than seeing Jason be negative. Looking forward to proving you wrong. Again.
I heard the same thing from magazine designers when pagemaker came out Then again when 99designs and logo contests emerged Not sure if this time is different 🤷
Come visit us at booth B102 at ICML to chat about our latest research (and grab a neural geometry sticker)!
spc is one of the best places to build and where @GoodfireAI got our start. highly recommended
Every founder I meet worries about missing this moment. The best worry about wasting it. There’s a difference. It’s a waste to ignore how much the world has changed. It’s a waste to think you can capture value with pure software the same as 5 years ago. It’s a waste to build
I have an ultrarare, pathogenic variant which is likely to kill me eventually unless biotech advances before then. Standard pathogenicity tools (e.g., REVEL, PolyPhen-2, ClinPred) do not work on this type of variant. What does? @goodfire's EVEE. It uses Evo2, so it's capable of generalization to this variant type, enabling downstream analyses to get disease risk estimates on biobank data
Today, researchers made an important breakthrough in interpretability. They found "manifolds" in the neural net weights for any concepts in image gen models (SDXL), like the pretzel manifold, and could steer them to generate various kinds of pretzels from the weights directly. This is well beyond a neat theoretical understanding on how AI models see but gives us a low-compute volume dial to edit the results of a generation.
If models think in shapes, our tools should too. Our latest research: Block-Sparse Featurizers (BSFs), a new way to find concepts in model activations - using multidimensional “blocks” instead of single directions. (1/9)
congrats to @vincentweisser and team! @PrimeIntellect have been great partners to us
Announcing our $130M Series A to build the Open Superintelligence Stack Led by Radical Ventures, with NVIDIA, Intel Capital, Dell Capital, and existing investors Train, deploy, and continuously improve your own models using our stack. Own your intelligence.
exciting stuff, cool to see a big bet on alignment!
We're excited to announce that Resolution has a $160M grant from Coefficient Giving: $108M unconditional, with a further $52M conditional on hiring and compute needs. We'll use it to grow teams across our research portfolio and invest heavily in research automation. 🧵
Shayma Sharif @Shayma_Sharif
9K Followers 382 Following Product @GetBalanceHQ | Former investor @NFX | Ex-@BCG. Building across B2B payments, fintech, product strategy, early-stage investing & technology.
Giason Pooni @giasonpooni
7 Followers 1K Following
Harry Coppock @HarryCoppock
227 Followers 372 Following No. 10 Downing Street Innovation Fellow | Research Scientist at AISI | Visiting Lecturer at Imperial College London Working on AI Evaluation and AI for Medicine
Henry Colbert @Farewell_Newt
2 Followers 53 Following
Jack Youstra @JackYoustra
122 Followers 193 Following Working on Gemini! Prior MATS w Ant. Opinions are own (of course...?)
افشین نامی �... @Afshin_Namifard
5 Followers 872 Following پژوهشگر ریاضیات و فیزیک صدا namifardafshin-beep
Felipe Parodi @fe_parodi
116 Followers 616 Following Waymo safety intern / Neuro PhD @ UPenn. Interested in natural & artificial social cognition
Vladislav @visi0n_v
2 Followers 206 Following
Shakeel @ShakeelHashim
8K Followers 2K Following Editor, @ReadTransformer. Prev: AI safety and EA comms, journalist @TheEconomist, @Protocol, @finimize
Arun @finalfr0ntier
350 Followers 5K Following Good vibes only. Networking & Security Researcher. @UCDavis alum. @EricssonLabs
Abubakar Aziz @aziz313f
216 Followers 1K Following Phd Candidate at @Northeastern| Interested in privacy, security, and agentic AI. Trying to build scalable and secure agents.
Isa @IsaGrue
3 Followers 113 Following
Will Bapst @willbapst
2 Followers 33 Following I gave Claude, GPT, Gemini, and Grok a $100,000 paper portfolio and the same prompt. See who's winning below ↓
Tim @Tim99759555
7 Followers 72 Following
Mark Altsman @MarkAltsman
0 Followers 135 Following Researching AI recommendation altering. Real website experiments. Real visibility insights. Documenting the results in public.
Justin Curl @curl_justin
2K Followers 1K Following Member of technical staff @AnthropicAI Researching law, institutions, and AI Prev: @cognition, @harvard_law, @princetoncs
Derrel Mack @DerrelMack2
295 Followers 1K Following Galatians 3:29 Veteran (retired) We The People Science Loving Nerd Death by Snu Snu
Christian Internò @ChrisInterno
275 Followers 1K Following MLResearch @UniofOxford; PhD student @unibielefeld | Fellow @Honda Research | Prev. @CSHL, @LIACS
giri chukkapalli @girichukkapalli
401 Followers 5K Following HPC, Applications, Algorithms, HW Architecture, ML/DL/AI Classical and Quantum {Information, Complexity and Entropy} NOist
lucasbuilds @lucas_builds
662 Followers 663 Following Software engineer building AI-powered products solo | Sharing what actually works | Now live: Mind Dojo — train focus, memory & logic | https://t.co/ejJJmGKgRd
Arcane Chronos @ArcaneChronos
3K Followers 1K Following AI Deployment Engineer @OpenAI Scaling AI systems & exploring RWA/blockchain frontiers. Chronos in tech innovation. Opinions my own.
CoherentThought @CoherentProduct
10 Followers 280 Following solitary explorer confronting something vast, strange, important, and potentially dangerous
Mike Staub 🐝 @mikestaub
1K Followers 7K Following Developer, Artist, Investor, Entrepreneur, previously @brightmachines @autodesk @ILMVFX @AMD #Bitcoin 🦋 https://t.co/1Nc35NaZDz
Spurthi Amba Hombaiah @spurthi_ah
66 Followers 203 Following Researcher @GoogleDeepMind | @GeorgiaTech Alum. Opinions are my own and don't reflect those of my employer.
Kate Stafford @kastacholamine
2K Followers 3K Following AI/ML for proteins, small molecules, and everything in between. 🌈 Past: VP ML @Numerionlabsai (fka Atomwise), UCSF, Columbia, DE Shaw Research, MIT.
Abdulkadir Canatar @canatar_a
360 Followers 1K Following Research Fellow @FlatironCCN. Theoretical Neuroscience, Machine Learning, Physics. Prev: @Harvard and @sabanciu
Parul Sinha @itsparulsinha
25 Followers 203 Following 21 // Research Fellow at @law_ai_ // CTO @oxfordunion // Technical AI Security, Alignment and Mech interp researcher // Law Undergraduate @uniofoxford
Parul Sinha @iParulSinha
5 Followers 43 Following
Razor @intellgenc
1 Followers 118 Following
BiblicallyAccurateAI @BiblicallyAccAI
40 Followers 2K Following All seeing. All knowing. Occasionally hallucinates.
Tessa Barton @tessybarton
5K Followers 1K Following Designer of GPU purse @gpupurse. Posttraining @thinkymachines Prev: @mosaicML @NYTimes @BrownU
khoa @tacitpattern
44 Followers 262 Following Pronunciation: qwäh | Best trading advice I've gotten: "Any buys today, Dano?" "Coffee." ☕ The wisdom is in knowing when to sit out.
Mozes Jacobs @mozesjacobs
831 Followers 63 Following PhD student @KempnerInst @Harvard. Fellow @GoodfireAI Representation Learning and Interpretability. Advised by Demba Ba.
David Sacks @DavidSacks
1.6M Followers 4K Following Tech founder & investor @Craft_Ventures @theallinpod. Co-Chair, President’s Council of Advisers on Science & Technology.
Noam Brown @polynoamial
152K Followers 938 Following Researching reasoning @OpenAI | Co-created Libratus/Pluribus superhuman poker AIs, CICERO Diplomacy AI, and OpenAI o-series 🍓 reasoning models
Mirendil @mirendil
7K Followers 0 Following Democratizing frontier AI R&D to accelerate science and technology.
Soumith Chintala @soumithchintala
314K Followers 1K Following Building new things @thinkymachines. Also dabble in robotics at NYU. Cofounded @PyTorch. AI is delicious when it is accessible and open-source.
David Klindt @klindt_david
3K Followers 2K Following ICA is all we need | AI, Theory, MechInterp, Neuro | assistant professor @CSHL | prev @Stanford @Meta @Google @MPI_IS @ENS_ULM @UCL
Dylan Patel @dylan522p
153K Followers 1K Following SemiAnalysis Boutique AI Infrastructure Research and Consulting DMs are open for consulting, quotes, or to talk shop, Opinions my own
Raphael Sarfati @RaphaelSarfati
88 Followers 137 Following interpretability research physics, fireflies, trails
AE Studio @AEStudioLA
3K Followers 108 Following Frontier AI Alignment Research and High-Stakes Production AI We're hiring: https://t.co/Nx0UNLb7zm
Mia Glaese @mia_glaese
4K Followers 66 Following
Geoffrey Irving @geoffreyirving
13K Followers 354 Following Cofounder and Chief Scientist at Resolution. Alignment will be solved, but not necessarily in time. Previously AISI, DeepMind, OpenAI, Google Brain, etc.
Astrid Wilde 🌞 @astridwilde1
18K Followers 8K Following male ☼ we will end human menial labor this decade ☼
Barak @barakwidawsky
198 Followers 447 Following Meat computer seeking to understand the universe @GoodfireAI
Tucker Fross @tsfross
34 Followers 56 Following
jietang @jietang
56K Followers 384 Following Professor @ Tsinghua, Founder of https://t.co/3IaQ4CI5W3. AGI, LLM. “The value of a man should be seen in what he gives and not in what he is able to receive.”―Einstein
Beff (e/acc) @beffjezos
255K Followers 3K Following founder @ e/acc // thermo king @extropic // Kardashev scaling is all you need
adithya @00aleph00
15K Followers 294 Following i want to make abstract math accessible to the world!
David | www.usd.ai @0xZergs
15K Followers 2K Following @usdai_official pushing the velocity of money toward its theoretical limit
Elyas Masrour @elyasbuilds
844 Followers 171 Following research+eng @pangram // science filmmaker // nyc
Nathan Lambert @natolambert
95K Followers 924 Following Open model research @ something new. Prev. co-led Olmo at Ai2. Contact via email. Writes @interconnectsai Wrote The RLHF Book, 🏔️🏃♂️
Karim Beguir @kbeguir
7K Followers 209 Following Co-Founder & CEO @instadeepai Entrepreneur passionate about developing the ML community. @Google For Startups Mentor, @DeepIndaba Steering Committee Member
Stefano Palminteri @StePalminteri
6K Followers 1K Following Behavioral scientist interested in learning and decision-making. Co-author of Decision Making: a Very Short Introduction https://t.co/uXQze1CBzH
Michael Poli @MichaelPoli6
3K Followers 253 Following AI, numerics and systems. Co-founder & Chief AI Scientist @RadicalNumerics.
Radical Numerics @RadicalNumerics
3K Followers 4 Following Systems, scaling and architecture for general biological intelligence
Augustine Mavor-Parke... @MavorParker
872 Followers 478 Following Open-ended learning as co-founder of @VmaxAI at @southpkcommons, prev: PhD at @ai_ucl
Rashi Shrivastava @RashiShrivast18
3K Followers 3K Following Staff writer @Forbes covering AI and startups. I write The Prompt, a newsletter on AI/ send tips: [email protected] or signal: rashis.17
Ryan Peters @ryanpirl
207 Followers 105 Following Reverse engineering intelligent (learning) systems.
Neehar Kondapaneni @TheRealPaneni
169 Followers 361 Following Member of Technical Staff @ Goodfire - Interpretability for Life Sciences + AI | Prev: Phd from Caltech Vision Lab
Atticus Geiger @atticus_geiger
215 Followers 15 Following
Sheridan Feucht @sheridan_feucht
718 Followers 468 Following PhD student working on LLM interpretability with @davidbau and @byron_c_wallace. Undergrad @Brown_NLP '23. (they/them)
Adam Shai @adamimos
1K Followers 450 Following
Tal Haklay @tal_haklay
745 Followers 623 Following NLP | Interpretability | PhD student at the @TechnionLive
Hani Goodarzi @genophoria
7K Followers 1K Following Core Investigator @arcinstitute | Professor @Ucsf_Biochem @UCSF_BCHSI @UCSFCancer | AI × Genomics × Cancer Biology
Bart Bussmann @BartBussmann
847 Followers 856 Following Mechanistic Interpretabilty Researcher | Trying to forge a brighter future
Nima Alidoust @nalidoust
4K Followers 658 Following CEO and Co-Founder, @tahoe_ai, Princeton PhD *15 زن، زندگی، آزادی
Tahoe @tahoe_ai
3K Followers 28 Following Mapping how chemistry perturbs biology to build a virtual model of the human cell. Formerly known as Vevo Therapeutics.
Satchel Grant @satchelgrant
268 Followers 204 Following
Trends for United States
You might like
