Cendekia @ EMNLP2026 | 🇮🇩 in 🇦🇪 @linearprobe
PhD student @mbzuai Interp @mint_nlp_mbzuai | Prev : MSc @mbzuai Research Intern @RIKEN_AIP EE @its_campus cendekia-airlangga.github.io Indonesia Joined June 2023-
Tweets844
-
Followers140
-
Following840
-
Likes9K
@machinyearning When yh
We revisited Independent Component Analysis (ICA), a decades-old statistical method, and found that it works far better for interpreting LLM activations than we expected! This old baseline deserves a second look. Across GPT-2 Small, Gemma 2 2B, and Qwen 3.5 9B Base, we find ICA is actually competitive with the public SAEs across interpretability, sparse reconstruction, concept prediction, and targeted probe perturbation, while fitting in minutes per layer and using 7–43× less parameter memory. We're releasing ICA Lens, which lets you quickly discover, annotate, and evaluate interpretable directions in LLM residual-stream activations. No need to train another dictionary! More broadly, our results suggest that non-Gaussianity is a useful inductive bias for efficiently finding interpretable directions. Try it: pip install icalens pypi.org/project/icalen… Paper: arxiv.org/abs/2606.11722… 1/8
We open source SEA-ASR-ID a fast speech-to-text for Bahasa Indonesian (Apache 2.0 364M params open model + 2,287.4 h speech dataset) - 1.5 s to transcript 79 minutes of Indonesian parliament (3,582× real time and 26–42x cheaper than Whisper large-v3 on 1x H100) - 5.5% WER with a 1B rescoring LM (on par with Whisper large-v3 1.5B), 6.1% WER with its n-gram LM (ties Whisper large-v3-turbo) Finetune on Google's EmbeddingGemma 2 inspire by @TslShahir @stablequan
What's the best resource to start learning about world models?
In reading group reviewing flow models, I realized I didn't understand the change-of-variables rule as fully as I probably should. So I put together this blog post and collection of visuals to help me more fully understand the Jacobian-inverse term. 💪 michaeljohncooper.com/writing/change…
Menurut saya indikasi negara maju itu salah satunya adalah ini: "Kita bisa neliti hal yang 'tidak berguna' dengan tenang" 'Tidak berguna' = hal yang, oleh awam, terlihat tidak bisa diaplikasikan langsung dan/atau belum ada gunanya saat ini. This applies to stem maupun soshum
Control engineernya sangar betul
super interesting to see anthropic's roadmap in 2021, i think they were very early in everything related to RL/tool use for the pre-training part, they mention "periodic batched backprop" and "skeletokenization" which might correspond to these two papers arxiv.org/abs/1711.10563 arxiv.org/abs/2105.09352
Anthropic released their 2021 research roadmap earlier this year (because of copyright litigation); I think some people might find it interesting: i.sm4.ca/concise-resear…
I wrote a blog post detailing the writing approach I took while working on my recent paper: sophielwang.com/blog/paper-wri… I describe the behind the scenes for how this paper came to be, why and how I storyboarded figures before running experiments, and some figure design tips! :)
Can “chicken” make a base model reason? 🐔 Yes! Why? With the right first tokens, base models can match RL in reasoning. We trace this to learned training data associations and use a simple data edit to make “chicken” elicit reasoning too. blog: sophielwang.com/cues
AI researchers are destroying themselves.
@GWirjawan Ditunggu episodenya dengan Demis Hassabis, Pak Gita. 👀
If doomscrolling X is part of your research workflow, we built something for you. Introducing Paperscrolling, now on iOS and Android 🚀 Follow the latest research with key ideas, figures, and audio explanations from alphaXiv Briefs Check it out here! play.google.com/store/apps/det…
Ramanujan wasn’t skipping proofs. He was hiding his chain of thought to prevent distillation. Hardy: “Show your work.” Ramanujan: “Sorry, reasoning tokens aren’t exposed via the API.”
@rayendito Change of variable, no?
Cool!
results for all claudes
Anshuman Lall @ NeurI... @anshuman1
900 Followers 4K Following LLM Post-Training/ Recursive Self-Improvement for Reasoning Co-founder/ MTS @ReasonCoreAI PhD @UCLA, Bachelors @iitbombay Led a team of ~500 PhD researchers
Arnau Marin-Llobet @Arnauya
1K Followers 2K Following PhD@harvard in comp neuro, interpretability. previously @AnthropicAI (stem) fellow, @Apple ...
Ardya Dipta Nandaviri... @ArdyaDipta
4K Followers 1K Following AI Consultant @amazon @aws | Alumni: @gojektech @CMU_Robotics @itbandung | Tweets are my own and not the view of my employer | ga accept akun profpic ga jelas
Athiya Deviyani @AthiyaD
12K Followers 2K Following PhD @LTIatCMU on trustworthy evals 🇮🇩 #letsgopens
Kaican Li @mjustlkc
30 Followers 236 Following PhD @HKUST, prev. UG @CUHKofficial | CV & NLP, Multimodal Reasoning, Self-Improving Agents, OOD Generalization
Siddhartha Bhattarai @siddhartha2028
38 Followers 2K Following LFX'26 Post Quantum Cryptography Alliance
Thanmay @yamnaht
457 Followers 179 Following MS Student @WSAI_IITM | @iitmadras | @AI4Bharat | @teamIvlabs
Earther @EartherAI
369 Followers 5K Following Researching Learning Systems | Deep Learning × Vision Mamba x Fusion Models
Vamshi Bonagiri @vamshibon
750 Followers 930 Following Alignment | prev CHAI @UCBerkeley | PhD @MBZUAI
Mahardika Krisna Ihsa... @mkrisnai
50 Followers 288 Following PhDing @MBZUAI & visiting student @sutdsg, continual yearner & yapper. Interested in principled method to make machine learn & generalize better . Ex-@wing_nus
佐藤魁/Kai Sato @KaiSatou0
62 Followers 136 Following A Doctor Student (D1) at Tohoku University (updated April 2025) / Member of Tohoku NLP Group (@tohoku_nlp)
Nitay Calderon @NitCal
475 Followers 796 Following Research Scientist @GoogleResearch | PhD @TechnionLive | NLP
Michael Hahn @mhahn29
2K Followers 1K Following LLM theory and interpretability. Professor at Saarland University @LstSaar @SIC_Saar. Previously PhD at Stanford @stanfordnlp. Hiring PhD students and Postdocs
Antonin Poché @Antonin_Poche
81 Followers 205 Following PhD Student doing XAI for NLP at @ANITI_Toulouse, IRIT, and IRT Saint Exupery. 🪄 Interpreto & Xplique libraries dev member. https://t.co/R75ZM3vrPh
Vincent Koc @vincent_koc
24K Followers 7K Following Futurist 🦄 Chief Architect @openclaw 🦞 Writer @forbes Tech Council Adjunct @MIT Ex @Microsoft @Qantas
JinYeong Bak @NoSyu
984 Followers 756 Following 박진영/JinYeong Bak/朴秦永 Leader @ https://t.co/fOvQXsEjo6 Conversation Modeling Researcher who is not good at talking
Kareem Elozeiri @elozeiri
27 Followers 183 Following NLP Master Student @MBZUAI , Doing Interpretability, model merging & MoEs.
ほりゆき @horiyuki42
2K Followers 931 Following Keio Mechanical Engineering (B2) & NLPer / Research Assistant in Robot Learning at UTokyo / ex-Sakana AI Applied Research Engineer / 19th ISO🥈
Smuggy @mamangsomaiy
96 Followers 446 Following
Emilio Villa Cueva @evllcv
80 Followers 271 Following PhD @ mbzuai. social intelligence, multimodality, & multilinguality
Jonibek Mansurov @M_Jonibek
39 Followers 135 Following
Dragono Flago @GirsangHansen
60 Followers 2K Following He loves the rainy night⛈️ The sage of his own tranquility🌬️🍃🕊️ Trying to memorize Eierschalensollbruchstellenverursacher🥚⛏
Tri Ahmad Irfan @irfan3
9K Followers 2K Following AI & Growth @airwallex. Built product with 1m+ users @ycombinator W22 • @twitter @X SF • @forbes 30u30
Jafar Isbarov @phylo_GENETIC
163 Followers 589 Following CS PhD student @VT_CS. Previously a visiting researcher @NYU_Courant. I work on AI Security, and Multilingual LLMs. Sapere aude.
Haruto Yoshida @yoshida_NLP
306 Followers 418 Following PhD Student @ Tohoku University (@tohoku_nlp) / NLP · Multimodal AI · Vision & Language / Interested in (understanding | generating) diagrams
Edd @erla_ndpg
290 Followers 2K Following Ex Computer Science student from Indonesia Only create X because AI Researchers are on X
Zrehaw @Zrehaw8059
25 Followers 1K Following
Bawaucor @Bawaucor207
31 Followers 992 Following
IkhlasulHanif @IkhlasulHanif0
339 Followers 608 Following AI Safety, NLP PhD student @mbzuai, currently visiting @sutdsg Prev: @univ_indonesia and JAIST Only create X because AI Researchers are on X
Feijiang Han😈 @feijianghan
572 Followers 805 Following CS PhD @UMDCS, working on LLMs and actionable interpretability. Prev @UpennNLP @Google RedNote: https://t.co/offEkZNZic Opinions are my own
Arnau Marin-Llobet @Arnauya
1K Followers 2K Following PhD@harvard in comp neuro, interpretability. previously @AnthropicAI (stem) fellow, @Apple ...
Jessica S. @obspeaking
225 Followers 176 Following Observationally speaking, im an astrophysicist and cosmologist. Most of the time want to fucked my computer.
Judea Pearl @yudapearl
85K Followers 401 Following Student of causal inference, human reasoning, and history of ideas, all viewed through the sharp lens of artificial intelligence.
Dasol Choi @NeurIPS'2... @solsol_choi
1K Followers 401 Following ex-fitness trainer → AI safety researcher leading safety evals @AIMSafety_ first author at NeurIPS/ACL/ECCV/AAAI
ellen livia ᯅ @ellen_in_sf
6K Followers 2K Following RL environment & Evals @SnorkelAI 👾🤖 forever student, AI research, art & life 🪂🇮🇩 i write @TDataScience & https://t.co/SXy0pMyyPv (2M+ reads)
Institute of Foundati... @IFM_AI
5K Followers 10 Following We create world-class foundation models that redefine what AI can do.
Kaican Li @mjustlkc
30 Followers 236 Following PhD @HKUST, prev. UG @CUHKofficial | CV & NLP, Multimodal Reasoning, Self-Improving Agents, OOD Generalization
SPAR @SPARexec
857 Followers 78 Following We're a part-time, virtual research program that gives students and early career professionals an opportunity to work with professional AI safety researchers.
Hua-Dong Xiong @HuaDongXiong
366 Followers 1K Following A restless mind interested in human and machine intelligence, indulged in forking paths in the garden.
Joanathan McIntosh @joanathanlmc
2K Followers 467 Following Co-Founder Billow (YCS26) | We're killing Deloitte | 2x Exited Founder | [email protected]
Mahardika Krisna Ihsa... @mkrisnai
50 Followers 288 Following PhDing @MBZUAI & visiting student @sutdsg, continual yearner & yapper. Interested in principled method to make machine learn & generalize better . Ex-@wing_nus
Lesmana Mandrakumara ... @lezmandrakumara
9K Followers 158 Following Antitesis Rakyat Jelata | Komedi, Opini dan Koleksi Pribadi
Hirokatsu Kataoka | �... @HirokatuKataoka
6K Followers 423 Following Chief Scientist @ AIST Academic Visitor @ Oxford VGG PI @ cvpaper.challenge PI @ LIMIT.Lab 3D ResNet (Top 0.5% in CVPR) FDSL (ACCV Award / BMVC Finalist)
Giang Nguyen @giangnguyen2412
815 Followers 212 Following By-design interpretability researcher @guidelabsai driven by a curiosity to understand the inner workings of AI and make it safer
Toshitake Asabuki @TAsabuki
364 Followers 149 Following Unit Leader @RIKEN_CBS | @RIKEN_PRI Investigating the computational principles of learning. Modeling synaptic plasticity and emergent neural dynamics.
Yasaman Bahri @yasamanbb
6K Followers 1K Following Research Scientist @GoogleDeepMind. AI x Physics. Ph.D. @UCBerkeley.
Jamie Simon @learning_mech
2K Followers 78 Following doing fundamental science of deep learning | PhD from Berkeley | can catch a whole egg in my mouth
Ryan Peters @ryanirl
422 Followers 164 Following reverse engineering intelligent (learning) systems. rl, evals, safety @emulated_ai
Vamshi Bonagiri @vamshibon
750 Followers 930 Following Alignment | prev CHAI @UCBerkeley | PhD @MBZUAI
Eric Ho @eric_ho
5K Followers 572 Following Co-Founder / CEO @GoodfireAI - AI interpretability research company
Richard Sutton @RichardSSutton
73K Followers 63 Following Student of mind and nature, libertarian, chess player, cancer survivor. @ https://t.co/pioQVVjSXz, UAlberta, amii, https://t.co/vPRUv44glx, The Royal Society, Turing Award
Weijie Su @weijie444
16K Followers 497 Following Training models @OpenAI, Professor @Wharton, CS, Math @Penn, coDir PRiML, PhD @Stanford
Alek Dimitriev @tensor_rotator
6K Followers 1K Following Member of Janitorial Staff at the Manhattan Project
David Sussillo @SussilloDavid
13K Followers 885 Following Neural reverse engineer. Author. Adjunct prof at Stanford. Previous: Meta Reality Labs, Google Brain Milton Hershey Alum.
Dyah Adila 🦄 @dyahadila_
336 Followers 256 Following Research Scientist @ScaledCognition | PhD from @WisconsinCS ; prev @GoogleResearch @awscloud
Azis S Prasetyotomo @azisseno
704 Followers 181 Following Software Engineer @ Meta · UW grad student I build AI and the interface Trying hard to get in shape
Kazuki Irie @kzkirie
600 Followers 211 Following Computer scientist @Yale. KI, der KI erschafft. Ex-farmer in 🇮🇹, ex-teacher in🇨🇭. All tweets are wrong, but some are useful. No need to follow back :-)
dutweets @dutweets
1.4M Followers 536 Following لنرفع علمنا بكل فخر 🇦🇪💙 #فخورين_بالإمارات #البعيد_يقرب
Emirates Support @EmiratesSupport
208K Followers 6 Following The official customer service channel for @emirates. We are here 24/7 to help you in English or Arabic.
Emirates @emirates
1.7M Followers 54 Following #FlyBetter. Need help? Send a comment or DM to @EmiratesSupport.
Etihad @etihad
843K Followers 27 Following The greatest journeys start with ambition. Beyond Borders. For customer support, send DM to @etihadhelp
Zayed International A... @AUH
91K Followers 392 Following Join our community of travel-lovers, and share the latest destinations, travel tips and exclusive airport offers!
Valentin Hofmann @vjhofmann
2K Followers 288 Following Assistant Professor @CisLMU @LMU_Muenchen | Formerly @allen_ai @uwnlp @UniofOxford @stanfordnlp @GoogleDeepMind
Nitay Calderon @NitCal
475 Followers 796 Following Research Scientist @GoogleResearch | PhD @TechnionLive | NLP
Elliot Arledge @elliotarledge
45K Followers 358 Following MTS @infinity_ai_ https://t.co/82V5vFob7t | made the 12 hr CUDA course | https://t.co/2hzezpo5Hi | @shipfr8 alum
Antonin Poché @Antonin_Poche
81 Followers 205 Following PhD Student doing XAI for NLP at @ANITI_Toulouse, IRIT, and IRT Saint Exupery. 🪄 Interpreto & Xplique libraries dev member. https://t.co/R75ZM3vrPh
Giri Kuncoro @girikuncoro
13K Followers 1K Following Kubernetes/Infra @SpaceXAI. Alumni: @tiktok @gojektech @vmware @cornell (tweets are my own)
Subham Sahoo @ssahoo_
4K Followers 91 Following Pioneering Diffusion LLMs | Team Lead @mbzuai - IFM | PhD @cornell
佐藤魁/Kai Sato @KaiSatou0
62 Followers 136 Following A Doctor Student (D1) at Tohoku University (updated April 2025) / Member of Tohoku NLP Group (@tohoku_nlp)
Callum McDougall @calsmcdougall
2K Followers 6 Following Interpretability Researcher @ Google DeepMind, ARENA founder. Guinness World Record holder for the largest number of clothes pegs held in one hand (yup, really)Trends for United States
You might like
