Honglak Lee

Can we train mobile GUI agents just by watching YouTube videos? Yes; we present MONDAY at #CVPR2025 — a large-scale dataset with 313K automatically annotated frames from 20K real-world mobile OS tutorials (iOS + Android). ✅ No manual annotation ✅ Robust cross-platform…

📣Thrilled to announce the drop of EXAONE 4.0, the next-generation hybrid AI. 🙌Prepare to be amazed by EXAONE’s capabilities. #EXAONE #LG_AI_Resrarch #HybridAI #AI lgresearch.ai/blog/view?seq=…

Last week, I gave a talk at Tsinghua University on scaling test-time compute with generative process verifiers/PRMs that verify reasoning with reasoning. @jietang thank you for the invite! TL;DR: A super data-efficient recipe to train generative process verifiers that can scale…

🚨Announcing SCALR @ COLM 2025 — Call for Papers!🚨 The 1st Workshop on Test-Time Scaling and Reasoning Models (SCALR) is coming to @COLM_conf in Montreal this October! This is the first workshop dedicated to this growing research area. 🌐 scalr-workshop.github.io

LLMs can reason step-by-step — but how do we verify each step? Process Reward Models (PRMs) can do step-level verification, but training them requires expensive, large-scale supervision. To improve efficiency, we introduce ThinkPRM: - Verifies reasoning via long CoT…

Excited to release EXAONE Deep, our new family of enhanced reasoning models excelling in Math, Science & Coding. EXAONE Deep 32B is competitive with several top models, and 7.8B & 2.4B rank #1 among comparable models. [model] huggingface.co/LGAI-EXAONE/EX… [blog]…

We are delighted to introduce NEXUS, an Agent AI system that tracks the lifecycle of training datasets used in AI models, comprehensively analyzes legal risks, and assesses potential threats related to dataset usage. NEXUS leverages our AutoCompliance agent to trace the full…

We are excited to present our paper on leveraging offline experiences for LLM agents using context-aware guidelines at #NeurIPS on Dec 11! 📖 AutoGuide: Automated Generation and Selection of Context-Aware Guidelines for Large Language Model Agents 🔗 arxiv.org/pdf/2403.08978

📝When training an LLM, we typically end up with substandard models: they perform best👍on some tasks but worse☹️on others. Should we discard🗑️these models? Well... If you Can't Use Them, Recycle Them: Optimizing Merging at Scale Mitigates Performance Tradeoffs 🧵👇 1/n

📣EXAONE 3.5-based 3 models are now available as open-source! Frontier-level Model, Top-tier Performance in Instruction Following and Long Context Capabilities. ☑️More Information on Our Blog : lgresearch.ai/blog/view?seq=… ☑️Model : huggingface.co/LGAI-EXAONE ☑️Technical Report :…

BREAKING NEWS The Royal Swedish Academy of Sciences has decided to award the 2024 #NobelPrize in Chemistry with one half to David Baker “for computational protein design” and the other half jointly to Demis Hassabis and John M. Jumper “for protein structure prediction.”

BREAKING NEWS The Royal Swedish Academy of Sciences has decided to award the 2024 #NobelPrize in Physics to John J. Hopfield and Geoffrey E. Hinton “for foundational discoveries and inventions that enable machine learning with artificial neural networks.”

Excited to share our work on leveraging code models to reason about action preconditions for sequential decision making tasks at NAACL 2024. I will be presenting the paper in the NLP Applications oral session on Wednesday morning. Come say hi! arxiv.org/abs/2311.09601

Very proud that our work on uncovering capabilities & limitations of LLMs on cultural commonsense has received the Social Impact Award at #NAACL2024--positive impact on society is what I'm personally striving for. Work led by S.Shen w/ @lajanugen M.Lee @honglaklee @soujanyaporia