🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

Excited to release my first lead project Magentic-UI at Microsoft Research, an OS web agent application designed for efficient human-agent interaction. CUA agents are cool but they're not so useful yet, Magentic-UI helps us study how to get value from them. github.com/microsoft/mage…

🚨 Lucie-Aimée Kaffee and I are looking for a junior collaborator to research the Open Model Ecosystem! 🤖 Ideally, someone w/ AI/ML background, who can help w/ annotation pipeline + analysis. docs.google.com/forms/d/e/1FAI…

What are NLP papers really saying about the purpose and use of their models/datasets? 🤔 Who are they for? What problems do they solve? How are they used? We built a framework + tool to: (1) analyze framing trends across papers (2) help authors reflect on their own framing 🧵

Can you train a performant language models without using unlicensed text? We are thrilled to announce the Common Pile v0.1, an 8TB dataset of openly licensed and public domain text. We train 7B models for 1T and 2T tokens and match the performance similar models like LLaMA 1&2

Do LLMs hallucinate randomly? Not quite. Our #ACL2025 (Main) paper shows that hallucinations under irrelevant contexts follow a systematic failure mode — revealing how LLMs generalize using abstract classes + context cues, albeit unreliably. 📎 Paper: arxiv.org/abs/2505.22630 1/n

Excited to share the results of my internship research with AI at Meta, as part of a larger world modeling release! What subtle shortcuts are VideoLLMs taking on spatio-temporal questions? And how can we instead curate shortcut-robust examples at a large-scale? Details 👇🔬

🔥Big exciting news - I've started a new company! 🚀 We are building AI agents that take actions in the real world by orchestrating the movement of physical goods. We're working with our first partners and are now growing the founding engineering team. We're building in NYC,

Thrilled to collaborate on the launch of 📚 CommonPile v0.1 📚 ! Introducing the largest openly-licensed LLM pretraining corpus (8 TB), led by Nikhil Kandpal Brian Lester Colin Raffel. 📜: arxiv.org/pdf/2506.05209 📚🤖 Data & models: huggingface.co/common-pile 1/

1) Thrilled to be at #Facct2025 for the first time this week, representing a meta-research paper on positionality statements at FAccT from 2018-2024, in collaboration with Solon Barocas (Solon Barocas) and Akshansh Pareek.

Theory of Mind (ToM) is crucial for next gen LLM Agents, yet current benchmarks suffer from multiple shortcomings. Enter 💽 Decrypto, an interactive benchmark for multi-agent reasoning and ToM in LLMs! Work done with Timon Willi & Jakob Foerster at AI at Meta & Foerster Lab for AI Research 🧵👇

A blizzard is raging in Montreal when your friend says “Wow, the weather is amazing!” Humans easily interpret irony, while LLMs struggle at it. We propose a 𝘳𝘩𝘦𝘵𝘰𝘳𝘪𝘤𝘢𝘭-𝘴𝘵𝘳𝘢𝘵𝘦𝘨𝘺-𝘢𝘸𝘢𝘳𝘦 probabilistic framework as a solution. arxiv.org/abs/2506.09301 @ #acl2025

🚨🚨 Excited to share a new paper led by Haiwen Li with the Community Notes team! LLMs will reshape the information ecosystem. Community Notes offers a promising model for keeping human judgment central but it's an open question how to best integrate LLMs. Thread👇

Extremely excited to share that I've joined the NLP department MBZUAI as a tenure-track Assistant Professor!

Existing AI Agent benchmarks are broken 🤖💔 Great work by Yuxuan Zhu and Daniel Kang identify + fix issues, and establish rigorous best practices for Agentic AI benchmarks! Check out the blog: ddkang.substack.com/p/ai-agent-ben…

🎉Personal update: I'm thrilled to announce that I'm joining Imperial College London Imperial College London as an Assistant Professor of Computing Imperial Computing starting January 2026. My future lab and I will continue to work on building better Generative Models 🤖, the hardest

Copyrighted 🚧, private 🛑, and sensitive ☢️ data remain major challenges for AI. FlexOlmo introduces an architectural mechanism to flexibly opt-in/opt-out segments of data in the training weights, **at inference time**. (Prior common solutions were to filter your data once

I'll be at #ICML2025 this week presenting SafeArena (Wednesday 11AM - 1:30PM in East Exhibition Hall E-701). Come by to chat with me about web agent safety (or anything else safety-related)!

I am speaking at 10 am PT on a slightly different topic than I usually talk about 🙂: "Simple Ideas Can Have Mighty Effects: Don't Take LLM Fundamentals for Granted" Check out if you're around. #ICML2025

Excited to present our AI Flaw Disclosure paper at #ICML2025 in Vancouver!🌲🌊🏔️ Swing by our poster session in East Exhibition Halls A-B E-606!