Zhiyuan

I can’t believe this jaw‑dropping comic was generated by GPT just by feeding it our paper directly🤯! It perfectly illustrates how meta‑ability training makes LRMs think better.

🚀Introducing GTA1 – our new GUI Agent that leads the OSWorld leaderboard with a 45.2% success rate, outperforming OpenAI's CUA! GTA1 improves two core components of GUI agents: Planning and Grounding. 🧠 Planning: A generic test-time scaling strategy that concurrently samples…

Customizing Your LLMs in seconds using prompts🥳! Excited to share our latest work with @HPCAILab, @VITAGroupUT, @k_schuerholt, @YangYou1991, @mmbronstein, @damianborth : Drag-and-Drop LLMs(DnD). 2 features: tuning-free, comparable or even better than full-shot tuning.(🧵1/8)

🚨🚨Reviewed around 20 papers for @ACMMM—but our own reviews were hidden & forced on us without expertise match. Time to rethink AI community peer review. 🤔 Our author team were assigned nearly 20 papers with no regard for our areas of expertise, received only a single round of…

🚀 Beyond “aha”: toward Meta‑Abilities Alignment! Zero human annotation enables LRMs masters strong reasoning abilities rather than aha emerging and generalize across math ⚙️, code 💻, science 🔬. Meta‑ability alignment lifts the ceiling of further domain‑RL—7B → 32B…

🚀 Beyond 'aha': toward Meta‑Abilities Alignment! By self‑synthesizes training tasks & self‑verifies rewards with zero human labels, LLM systematically masters core reasoning abilities rather than aha emerging and generalize across math ⚙️, code 💻, science 🔬. Meta‑ability…

Although the ICLR main conference is coming to an end, we are excited to invite you to the Reasoning and Planning for LLMs Workshop, which will be held all day on Monday, April 28. We are honored to host an outstanding lineup of keynote speakers and panelists from Meta, OpenAI,…

Welcome to use JudgeLRM! Compare any Hugging Face language models by asking your own questions, and explore JudgeLRM’s reasoning and detailed comparisons! Demo: huggingface.co/spaces/nuojohn… Paper: huggingface.co/papers/2504.00… Model: huggingface.co/nuojohnchen/Ju… Code: github.com/NuoJohnChen/Ju… We…

🚀 Exciting news! The ICLR 2025 LLM Reasoning & Planning Workshop is offering several Student Registration Grants to support early-career researchers 💡 Free ICLR registration for in-person full-time students! Apply by March 2, 2025. More info: …shop-llm-reasoning-planning.github.io Submit…

🚀 Call for Reviewers! 🚀 Our Workshop on Reasoning and Planning for LLMs at ICLR 2025 @iclr_conf has received an overwhelming number of submissions! We are looking for reviewers to help ensure a high-quality selection process. 🔹 Max 2 papers per reviewer 🔹 Review deadline:…

We are excited to announce that our workshop will be held on April 28 in Singapore. Due to numerous requests for extensions, we have decided to extend the submission deadline by 4 days to February 6 (AoE). We look forward to receiving your submissions and can't wait to see you at…

🚀 Excited to announce our World Models: Understanding, Modelling and Scaling Workshop at #ICLR2025! 🎉 Keynote speakers, panellists, and submission guidelines are live now! Check out: 👉 sites.google.com/view/worldmode… Join us as we explore World Understanding, Sequential Modelling,…

Our poster presentation at #NeurIPS2024 will take place today from 11:00 AM to 2:00 PM in West Ballroom A-D, Poster #7004. We warmly welcome you to stop by and engage with us!