johnny @johnnylin
@neuronpedia someone told me you're supposed to reply to all the replies to yoru tweets to boost ur twitter rank or something. im not gonna do that san francisco Joined January 2009-
Tweets24
-
Followers549
-
Following4
-
Likes56
🔢➡️🔤
New Anthropic research: Natural Language Autoencoders. Models like Claude talk in words but think in numbers. The numbers—called activations—encode Claude’s thoughts, but not in a language we can read. Here, we train Claude to translate its activations into human-readable text.
hmm i should have used a better example for the first post in that tweet. llama confabulates a bit here and people are semi-rightfully arguing that llama is just confused, not lying. here's a better one where llama's reasoning is more clear: neuronpedia.org/nla/cmow97q1r0…
An average person can't look a CT scan and identify cancer, but radiologists can. An average person can't look at Llama's model activations and identify lying, but Natural Language Autoencoders sometimes can. Here, an activation verbalizer shows Llama planning to lie. 🧵
Researchers can use the Neuronpedia interactive interface here: neuronpedia.org/gemma-2-2b/gra… And we’ve provided an annotated walkthrough: github.com/safety-researc… This project was led by participants in our Anthropic Fellows program, in collaboration with Decode Research.
Announcement: we're open sourcing Neuronpedia! 🚀 This includes all our mech interp tools: the interpretability API, steering, UI, inference, autointerp, search, plus 4 TB of data - cited by 35+ research papers and used by 50+ write-ups. What you can do with OSS Neuronpedia: 🧵
Neuronpedia now hosts Chain-of-Thought! Steer and inspect Deepseek-R1-Distill-Llama-8B with SAEs trained by @Open_MOSS on @neuronpedia (linked below). One fun initial result: the model can easily be steered into "overthinking/anxious" mode with a single latent.
Google DeepMind has a new way to look inside an AI’s “mind” trib.al/UN5W3Uh
Gemma Scope allows us to study how features evolve throughout the model and interact to create more complex ones. Want to learn more? Here’s an interactive demo made by @neuronpedia - no coding necessary ↓ dpmd.ai/gemma-scope
Want to learn more? @neuronpedia have made a gorgeous interactive demo walking you through what Sparse Autoencoders are, and what Gemma Scope can do. If this could happen pre-launch, I'm excited to see what the community will do with Gemma Scope now! neuronpedia.org/gemma-scope
Sparse Autoencoders act like a microscope for AI internals. They're a powerful tool for interpretability, but training costs limit research Announcing Gemma Scope: An open suite of SAEs on every layer & sublayer of Gemma 2 2B & 9B! We hope to enable even more ambitious work
exciting new research from @apolloaisafety and @jordantensor: E2E SAEs (w/ ~700k features) are now live on @neuronpedia - the first to use dual UMAPs for visual comparison and exploration between SAE training methods. check it out at neuronpedia.org/gpt2sm-apollojt
Proud to share Apollo Research's first interpretability paper! In collaboration w @JordanTensor! ⤵️ publications.apolloresearch.ai/end_to_end_spa… Identifying Functionally Important Features with End-to-End Sparse Dictionary Learning Our SAEs explain significantly more performance than before! 1/
Terrific work by @saprmarks and team! 🥳 We really enjoyed working with them to get their Sparse Autoencoders onto @neuronpedia. You can explore, search, and test their 622,594 features here: neuronpedia.org/p70d-sm
Can we understand & edit unanticipated mechanisms in LMs? We introduce sparse feature circuits, & use them to explain LM behaviors, discover & fix LM bugs, & build an automated interpretability pipeline! Preprint w/ @can_rager, @ericjmichaud_, @boknilev, @davidbau, @amuuueller
6/ Oh and of course, @neuronpedia is publicly available for anyone to experiment and play with at neuronpedia.org. Let us know what you think!
5/ Thanks to @JBloomAus for support, @NeelNanda5 for TransformerLens, @ch402 @nickcammarata for inspiration from OpenAI Microscope, and William Saunders for Neuron Viewer. It's time to accelerate (interpretability research). 🚀🔬 lesswrong.com/posts/BaEQoxHh…
1/ Introducing Neuronpedia: an open platform for interpretability research with hosting, visualizations, and tooling for Sparse Autoencoders (SAEs). Let's try it out! ➡️ Neuronpedia lets us instantly test activations of SAE features with custom text. Here's a Star Wars feature:
Super impressed by @johnnylin's Interactive Interface for exploring my GPT2 Small SAE Features. neuronpedia.org/gpt2-small/res…. First 5000 for each layer are there with the rest coming shortly! We've updated the feature-activation highlighting to better show multiple fires per context!
best IoT feature: devices that automatically update for daylight savings time
Nikolai Zhukov @Zhukovi
53 Followers 482 Following Human. Researcher. Math, software engineering, visualization of complex systems and data. Research together with production in machine learning.💻❤️🏓🎨 🔦🎶
LearnAI @learnAI_BeFree
15 Followers 1K Following
Marco Mascorro @Mascobot
18K Followers 3K Following Researching LLMs | Roboticist | Prev: Partner @a16z (inv @cursor_ai, @thinkymachines, @bfl_ml, etc), cofounded Fellow Robots, @singularityu, @BMW Research
Suhail Nadaf @Suhailnadaf509
31 Followers 178 Following independent mech interp researcher · FOSS advocate
Tonny Ouma @techie4real
107 Followers 312 Following Applied AI @AWS. Cloud enthusiast. Opinions are my own.⚒️ blue check
Joesph Mcmillan @crazysMusics
190 Followers 2K Following House/Techno with feeling. Producer, Engineer, Songwriter, MusicTech Lecturer. Antiracist. Trans ally
Alix (Mini) @alixwiessser
1K Followers 2K Following phd in modern nonsense • https://t.co/Gmxr30zVz4
Theophilus Ige @TheOphige
340 Followers 4K Following ML Research + Blockchain Protocol + ZK =? DeAI 🦀
unruly abstractions @unrulyabstract
10 Followers 978 Following https://t.co/Gwjhi1Sfma all my failures are hopefully interesting
HK Tan @maxtanhk
0 Followers 30 Following
Mateusz G @MateuszG14
2 Followers 65 Following
Marçal Comajoan Cara @marcal_cc
60 Followers 676 Following phd student @berkeley_ai! research in ai for autonomous scientific discovery and to advance genetics
beyondr🏴☠️�... @thebeyondr
286 Followers 2K Following nah, I'd adapt. building tools powered by models merc @raidguild prev @privacyethereum
ZioN @ZionSattwa
6 Followers 78 Following
ManuLion.eth @ManuLionCrypto
5K Followers 4K Following AI, Robotics, Space, Quantum, Tech, Crypto, Web3 🦁
Tyler Sorensen @tylerya
376 Followers 1K Following
Rishub Jain @shubadubadub
453 Followers 519 Following Research Engineer at @GoogleDeepMind, currently working on AI Safety & Responsibility
Jatin Kashyap @JayKayNJIT
40 Followers 155 Following Atomic Models of 2D/Bio-Materials | #1stGen Scientist | #Quantum Compute | #Neuromorphic Intelligence | #CompChem #DrugDesign #MaterialsScience | #PiNN #DFT #MD
tmc/acc @traviscline
10K Followers 2K Following Hacking on mlx-go, it2, langchaingo, nlm, cgpt, OS AI stuff. 💼 ai. prev Facebook, Alto, Virta, Brev, NVIDIA, current: solo
Scott Kerr @ScottDangerKerr
499 Followers 2K Following Teaching Tech Skills at https://t.co/LKJdCIpxwo. Interested in research on AI and space.
Karmabir Chakraborty @karmabirr
4 Followers 113 Following
Wynand Jordaan @WynandFJordaan
4 Followers 71 Following
Garuna @garuna_athem
40 Followers 533 Following Deliberately stupid | Share a smile and witness the world reflecting joy
Thisisbrigette_b. @BridgetteB5059
96 Followers 2K Following Am a tall fashioned lady with (5'8"), buxom, and shapely blonde bombshell also a spanish descent.
beriox @beri0x
8 Followers 329 Following
Aruna S @arunasank
363 Followers 751 Following Curious about model cognition! EECS PhD student @MIT_CSAIL Past @medialab @hotosm @outreachy
Govind K @t2govind
2K Followers 7K Following Research Engineer @Microsoft. Developing experiences that people want
Richard H @0xRichardH
117 Followers 2K Following (Learning & Buidling) Stay Hungry. Stay Foolish. Keep Exploring.
Mike Staub 🐝 @mikestaub
1K Followers 7K Following Developer, Artist, Investor, Entrepreneur, previously @brightmachines @autodesk @ILMVFX @AMD #Bitcoin 🦋 https://t.co/1Nc35NaZDz
Aaron Bergman 🔍 @AaronBergman18
5K Followers 2K Following 👎: suffering | 👍: EA, AI alignment, decoupling, R, cringe, amateur pharmacology + programming | Georgetown ‘22 (math+econ+phil) | Career status: 🤷♂️
Lyhehe @Lyhehe1
5 Followers 364 Following
Aaron Bergman 🔍 @AaronBergman18
5K Followers 2K Following 👎: suffering | 👍: EA, AI alignment, decoupling, R, cringe, amateur pharmacology + programming | Georgetown ‘22 (math+econ+phil) | Career status: 🤷♂️
The Office on Peacock @theofficetv
1.2M Followers 220 Following Peacock Person's Paper People. #TheOffice is streaming now on @Peacock.You might like
