Been Kim

Research Scientist at Google DeepMind, PhD from MIT. Make machines empower people.

[Pulkit Verma @pulkit_verma]

The program for the #ICLR2026 Workshop "From Human Cognition to AI Reasoning" is now available. We have a fantastic lineup of talks.

🔗 https://t.co/InLGHBSR9I

Invited Speakers: @DocRachidAlami, @_beenkim, @ced_zhang

Co-Organizers: @julie_a_shah, @sarath_ssreedh, @si_tulli

[Been Kim @_beenkim]

Prompt engineering is still a black box. Why does changing X drastically change Y? Are there governing rules behind this evolution? Our new work proposes a simple way to uncover factors that might matter when refining prompts 👇

[Neha Kalibhat @NehaKalibhat]

Thrilled to share that our paper on "Interpreting and Controlling Model Behavior via Constitutions for Atomic Concept Edits" has been accepted at AISTATS 2026! 🚀🚀
Read more about how input mutations can be mapped to interpretable behavioral insights.
https://t.co/iRPRJoyAso
🧵

[Been Kim @_beenkim]

@francoisfleuret As my brain is taken over by “wrongfully accused” sentiment, I typed in my password to their website, brainlessly…

[Been Kim @_beenkim]

I got my account back! Thank you, first and foremost, to everyone—friends, GDM colleagues---who personally alerted me to this incident and retweeted that I'm hacked, as well as folks at X who helped me regain access. While this incident was terrible (I heard the scammers made huge money out of this), I feel incredibly lucky to have folks who cared♥️♥️♥️ (details of how this happened 👇)

[Been Kim @_beenkim]

@savvyRL haha What do you think? Do I really sound like Been Kim now? :) even if you text me, it's possible that scammers has my phone too. even if we can GVC, it might be an AI-generated content. We just have to meet in person, since robotics is not quite there yet. 😵‍💫😵‍💫😵‍💫

[Been Kim @_beenkim]

@ashVaswani Congrats!! 🎉

[Susan Zhang @suchenzang]

What a privilege it is to have time as your most valuable currency.

[Christopher Potts @ChrisGPotts]

Safety-oriented interpretability researchers should be focused on AI systems, not individual model artifacts. A snippet from the NeurIPS CogInterp workshop panel on Sunday:

[Christopher Potts @ChrisGPotts]

This post seems to describe substantially the same view that I offer here:

https://t.co/LoNw7jFltD

Why are people describing the GDM post as concluding that mech-interp is a failed project? Is it the renaming of the field and constant talk of "pivoting"?

[Been Kim @_beenkim]

@giffmana Wait until he says six seven. 🙃

[Been Kim @_beenkim]

Tomorrow 9:30am #NeurIPS2025 Room 30A-E I'll talk about " 📈Towards Pareto frontier of interpretability:
15 years of interpretability research in 15 mins"🚅
@ mech interp workshop https://t.co/p3Hi5PV08V

[Been Kim @_beenkim]

@doomie @shaneguML go harass Dumi 🤣

[Been Kim @_beenkim]

Take that @doomie Samy Bengio! Hehehe

[Been Kim @_beenkim]

Our work out there in the wild 🥹

[Zi Wang, Ph.D. @ziwphd]

🔥 Proactive Co-Creator is officially LIVE in @GoogleAIStudio!
Stop guessing prompts. Start collaborating. Use it now to remix ideas and generate images, stories, and video with an AI that proactively helps you create.

🔗 Try it here: https://t.co/wPQrIpEglC

📍 At #NeurIPS2025? Come see the live demo TODAY (Dec 3) 9AM - 1:30PM | Google Booth #1533 (Kiosk 3)

🧠 Our research @GoogleDeepMind : We’re turning theory into practice. Read the papers behind the tech:
Concept Edits (Tech Report): https://t.co/5h3IASHJAu
Proactive Agents (ICML 25'): https://t.co/j7eWxjZL3j
QuestBench (NeurIPS 25'): https://t.co/E5jVYVqaxJ

[Zi Wang, Ph.D. @ziwphd]

🔥 Proactive Co-Creator is officially LIVE in @GoogleAIStudio!
Stop guessing prompts. Start collaborating. Use it now to remix ideas and generate images, stories, and video with an AI that proactively helps you create.

🔗 Try it here: https://t.co/wPQrIpEglC

📍 At #NeurIPS2025? Come see the live demo TODAY (Dec 3) 9AM - 1:30PM | Google Booth #1533 (Kiosk 3)

🧠 Our research @GoogleDeepMind : We’re turning theory into practice. Read the papers behind the tech:
Concept Edits (Tech Report): https://t.co/5h3IASHJAu
Proactive Agents (ICML 25'): https://t.co/j7eWxjZL3j
QuestBench (NeurIPS 25'): https://t.co/E5jVYVqaxJ

[Stanford NLP Group @stanfordnlp]

Awesome @NeurIPSConf keynote this morning by @YejinChoinka on The Art of (Artificial) Reasoning – and her broader thoughts and wishes on the future of Artificial Intelligence

https://t.co/Zn5y7LWOV1

[Been Kim @_beenkim]

Add: 9:30am on Sunday at Neurips, i'll touch upon this at the mech interp workshop keynote https://t.co/MEoMMFgKRB

[Been Kim @_beenkim]

1/8 Pareto Frontier 🤠for Human-centered AI 📈: We all want to build AI that is good for humans, but the path is often paralyzed by complexity. Either “oh my god, it’s too complicated😱” or delusional “I have a warm and fuzzy feeling of understanding 🥴”? "It’s hard because it depends.🤷" is the enemy of progress. We need a Pareto Frontier for Human-centered AI. 🧵👇

[Been Kim @_beenkim]

8/8 Making AI benefit humans takes a village. 🌍 But a village needs a shared language. Let's stop guessing and start measuring the frontier.📷

a short write-up: https://t.co/Zg7LotVcVI