🍓 there's a lot of o1 hype on my feed, so i'm worried that it might be setting the wrong expectations what o1 is: the first reasoning model that shines in really hard tasks, and it'll only get better. (i'm personally psyched about the model's potential & trajectory!) what o1

I can't believe the next generation of #Wikipedia #NLProc researchers won't waste endless nights parsing Wikitext anymore! New structured Wikipedia dataset in English & French on Hugging Face! Say goodbye to manual parsing and hello to machine-readable content 🤖👇

It's a blockbuster performance by our lab LCS2 Lab at #EMNLP2024. 8 out of 8 papers have been accepted. Our gang is coming to Miami. 👊 EMNLP 2025 IIT Delhi

🚀 Introducing an open-access course on #LLMs: Intro & Recent Advances! 🌐 - Covered both traditional and recent research in LLMs - A collaborative effort between industry and academia - W𝐞𝐛𝐬𝐢𝐭𝐞: lnkd.in/g_GEywQF - 𝐋𝐞𝐜𝐭𝐮𝐫𝐞 𝐯𝐢𝐝𝐞𝐨𝐬: lnkd.in/gGtmp8vP

📢 Excited to introduce CBT-Bench (huggingface.co/papers/2410.13…), a new benchmark that systematically evaluates LLMs’ capabilities in Cognitive Behavioral Therapy (CBT) across three Levels:

The AI community has been lacking models that are: - multilingual 🗣️ - multimodal 🖼️ - multicultural 🌎🌍🌏 Our new paper introduces - Pangea, a model - PangeaInstruct, a dataset - PangeaBench, an eval benchmark towards this goal! neulab.github.io/Pangea/

I will be in Miami this week for EMNLP '24! I’ll be presenting our latest work in my PhD and would like to discuss career opportunities (industry/academia). Let’s connect, exchange ideas, and make the most of this #EMNLP.

Are you ready for an early Christmas present from our team at Microsoft Research? Introducing the most powerful smol model ever built in the world! Welcome to Phi-4! 👇

I am honoured to receive the prestigious Humboldt fellowship.

🌟 𝐀 𝐍𝐞𝐰 T𝐞𝐱𝐭𝐛𝐨𝐨𝐤 -- 𝐈𝐧𝐭𝐫𝐨𝐝𝐮𝐜𝐭𝐢𝐨𝐧 𝐭𝐨 𝐋𝐚𝐫𝐠𝐞 𝐋𝐚𝐧𝐠𝐮𝐚𝐠𝐞 𝐌𝐨𝐝𝐞𝐥𝐬 🌟 I am excited to share the release of my new textbook, 𝘐𝘯𝘵𝘳𝘰𝘥𝘶𝘤𝘵𝘪𝘰𝘯 𝘵𝘰 𝘓𝘢𝘳𝘨𝘦 𝘓𝘢𝘯𝘨𝘶𝘢𝘨𝘦 𝘔𝘰𝘥𝘦𝘭𝘴 (#LLMs) -- Perhaps the first textbook on LLMs.

#WWW2025 | Our work has been accepted for oral presentation at TheWebConf 2025 Title: Figurative-cum-Commonsense Knowledge Infusion for Multimodal Mental Health Meme Classification Stay tuned for preprint, data, and code on our lab's webpage: FLaMe Research Lab 1/n

Kicking off the year with a bang -- 4 papers accepted in prestigious venues this month! #ICLR2025 -- 𝐋𝐋𝐌 𝐜𝐨𝐦𝐩𝐫𝐞𝐬𝐬𝐢𝐨𝐧: We introduce 𝐏𝐫𝐮𝐧𝐞𝐍𝐞𝐭, a novel, dataset-free policy learning approach to model pruning, achieving high compression efficiency and

🎉 Excited to share that our work has been accepted to #Findings of NAACL 2025! 📜 Title: Target-Augmented Shared Fusion-based Multimodal Sarcasm Explanation Generation 👥 Authors: Palaash Goel, Dushyant Singh Chauhan, Md Shad Akhtar #NAACL2025 #NLP #Multimodal #AIResearch

#CVPR2025 Area Chairs (ACs) identified a number of highly irresponsible reviewers, those who either abandoned the review process entirely or submitted egregiously low-quality reviews, including some generated by large language models (LLMs). 1/2

Following a thorough investigation, the Program Chairs (PCs) decided to desk-reject 19 papers authored by confirmed highly irresponsible reviewers, which would have been accepted otherwise, in accordance with the previously communicated CVPR 2025 policies. 2/2

Our #ICML25 Paper "𝐄𝐧𝐨𝐮𝐠𝐡 𝐨𝐟 𝐒𝐜𝐚𝐥𝐢𝐧𝐠 𝐋𝐋𝐌𝐬! 𝐋𝐞𝐭'𝐬 𝐅𝐨𝐜𝐮𝐬 𝐨𝐧 𝐃𝐨𝐰𝐧𝐬𝐜𝐚𝐥𝐢𝐧𝐠" challenges the obsession with scaling laws -- it's time to #downscale LLMs for efficiency, sustainability & real-world usability. Remember: SLMs are the future! Kudos