New paper: How can you tell if a transformer has the right world model?
We trained a transformer to predict directions for NYC taxi rides. The model was good. It could find shortest paths between new points
But had it built a map of NYC? We reconstructed its map and found this:
906 Followers 427 FollowingAsst. Professor @Rutgers. Previously, privacy researcher @Apple, postdoc at @JohnsHopkins, @PennState, and a PhD student @UWaterloo
5K Followers 2K FollowingIncoming professor @UCBerkeley in EECS and Statistics.
Principal Researcher @ Microsoft Research.
AI & RL. Lecture Notes: https://t.co/bhgL3aKIk0
78 Followers 62 FollowingWe build bicycles for the mind — algorithms that enhance rather than replace human capacity. Economists + computer scientists @MIT + @UChicago.
1K Followers 1K FollowingAssistant Professor at @WisconsinCS. Machine learning, privacy, and memorization. Postdoc @uwcse and PhD at Boston University.
281 Followers 1K FollowingSoftware Engineer • Teaching CS & building communities • Sharing my own thoughts here • Solutions Eng @tiriig_tech @beerooleApp • Program @fikrcamp
207 Followers 2K Following49k followers on LinkedIn | 41k on YouTube | 6k on Instagram | 6k newsletter subscribers | Get real-life superpowers with AI!! Follow me for the hottest posts!!
582 Followers 1K FollowingEconomist
Postdoctoral Researcher @UCStoneCenter @UChicago
Ph.D. in Economics from @Penn
Formerly: ⚽️ Swiss National Team
#FirstGen
24K Followers 10K Followingcognizing structures of information processing systems, in all their forms | applied category theory | the Wisdom Basin hypothesis | cancel heat death
933 Followers 217 Followingex-Member of Technical Staff @ OpenAI | ex-CTO @ Pyte | ex-Researcher @ Microsoft | PhD @ MIT | Gold Medal @ ICPC World Finals | 2 x Gold Medal @ IOI
1K Followers 353 FollowingProfessor of Computer Science. Works on Theoretical Computer Science, Algorithmic Game Theory, and Multi-Agent Interaction in Markets and Games.
906 Followers 427 FollowingAsst. Professor @Rutgers. Previously, privacy researcher @Apple, postdoc at @JohnsHopkins, @PennState, and a PhD student @UWaterloo
53K Followers 674 FollowingProfessor, Santa Fe Institute. Mostly posting on https://t.co/4NpA2IL5Va (at-melaniemitchell). More thoughts at https://t.co/nC43NHRozX.
651 Followers 2K Followingasst prof at @gatech_scs
FODSI Postdoctoral Fellow @MIT
PhD from @Berkeley_EECS;
Ex: Microsoft Research, Apple
Apple AI/ML Research Fellow 2023
5K Followers 312 FollowingAssociate Professor @Berkeley_EECS. NeurIPS 2026 Program co-Chair. Research on Foundations of ML and AI. Sloan and Schmidt AI2050 fellow.
11K Followers 60 FollowingSolving Machine Learning at Microsoft in New York.
https://t.co/ZpdQV4IsHY pandemic past president.
https://t.co/MkluiHpWF7 makes RL real.
https://t.co/wK8xQaQGwf for thinking out loud.
29K Followers 505 FollowingHelping the world prepare for powerful AI. Risk assessment @METR_evals (opinions my own). Blogs: Planned Obsolescence (AI), Good Bones (whatever's on my mind).
78K Followers 468 FollowingProfessor @ 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
7K Followers 59 FollowingCryptographer (lattices/post-quantum), Professor @UMichCSE, CS(cientific)O @AlgorandFoundation, PhD @MIT_CSAIL. Previously @gatech_scs. Here I speak for myself.
53K Followers 230 Followingidiot. cuda og, harvard val, morgan prize, society of fellows, 1 hilbert problem so far, creating friendly, SAFE, delightful, supergenius ..things @anthropicai
65K Followers 947 FollowingAssistant professor (of mathematics) at the University of Toronto, visiting Harvard Fall 2026. Forever confused. Algebraic geometry, number theory, etc. He/him.
78 Followers 62 FollowingWe build bicycles for the mind — algorithms that enhance rather than replace human capacity. Economists + computer scientists @MIT + @UChicago.
5K Followers 2K FollowingIncoming professor @UCBerkeley in EECS and Statistics.
Principal Researcher @ Microsoft Research.
AI & RL. Lecture Notes: https://t.co/bhgL3aKIk0
95K Followers 178 FollowingProf @the_IAS. On leave @Columbia. Head of Research @a16zcrypto. Wrote Algorithms Illuminated, 20 Lectures on Algorithmic Game Theory, etc.