Excited to share our new survey on the reasoning paradigm shift from "Think with Text" to "Think with Image"! 🧠🖼️ Our work offers a roadmap for more powerful & aligned AI. 🚀 📜 Paper: arxiv.org/pdf/2506.23918 ⭐ GitHub (400+🌟): github.com/zhaochen0110/A…

🧠 How can AI evolve from statically 𝘵𝘩𝘪𝘯𝘬𝘪𝘯𝘨 𝘢𝘣𝘰𝘶𝘵 𝘪𝘮𝘢𝘨𝘦𝘴 → dynamically 𝘵𝘩𝘪𝘯𝘬𝘪𝘯𝘨 𝘸𝘪𝘵𝘩 𝘪𝘮𝘢𝘨𝘦𝘴 as cognitive workspaces, similar to the human mental sketchpad? 🔍 What’s the 𝗿𝗲𝘀𝗲𝗮𝗿𝗰𝗵 𝗿𝗼𝗮𝗱𝗺𝗮𝗽 from tool-use → programmatic

I was saddened to find that out of the six papers I submitted, the reviewer response rate was 2/20, and out of the 12 papers I reviewed, the reviewer response rate was 16/43 (12 of which were my own)😢 #ARR #emnlp ACLRollingReview

Conference on Language Modeling decisions are out, and so are we The strength of submissions this year amazed us! Many many hard decisions 😩 + Aditi Raghunathan, Eunsol Choi, Ranjay Krishna 😴😴😴

Check out our new paper: “How Do Vision-Language Models Process Conflicting Information Across Modalities?”! Vision-language models often struggle with conflicting inputs - we show how their internal representations and key attention heads reveal when and how this happens, and

COLM 2025 accepted submissions are now public: openreview.net/group?id=colmw… Congratulations to all the authors, and see you all in Montreal!

Our department (ECE) at Princeton is hiring in AI this year!📢 Please consider applying and joining us: princeton.edu/acad-positions…

So many disappeared Google Scholar citations 😢

You can now train Vision LLMs with Reinforcement Learning in our free notebook! Unsloth VLM RL via GRPO: 1.5× faster, 90% less VRAM, 15× longer context & no accuracy loss. Guide: docs.unsloth.ai/new/vision-rei… GitHub: github.com/unslothai/unsl… Qwen2.5-VL Colab: colab.research.google.com/github/unsloth…

Are there conceptual directions in VLMs that transcend modality? Check out our COLM spotlight🔦 paper! We analyze how linear concepts interact with multimodality in VLM embeddings using SAEs with Chloe H. Su, @napoolar, Sham Kakade and Stephanie Gil arxiv.org/abs/2504.11695

One paper accepted at NeurIPS2025😊. San Diego I’m coming!

🚀 Introducing DeepSeek-V3.2-Exp — our latest experimental model! ✨ Built on V3.1-Terminus, it debuts DeepSeek Sparse Attention(DSA) for faster, more efficient training & inference on long context. 👉 Now live on App, Web, and API. 💰 API prices cut by 50%+! 1/n