Excited to share our new work led by Kunal Jha : scaling training to more diverse environments is key to human-AI cooperation!

Sampler is crucial for faster convergence of online DPO! Check out out paper: arxiv.org/abs/2409.19605 #ICLR2025

🧠 Your LLM should model how you think, not reduce you to preassigned traits 📢 Introducing LoRe: a low-rank reward modeling framework for personalized RLHF ❌ Demographic grouping/handcrafted traits ✅ Infers implicit preferences ✅ Few-shot adaptation 📄 arxiv.org/abs/2504.14439

Excited to share our work led by Yiping Wang RLVR with only ONE training example can boost 37% accuracy on MATH500.

So excited to announce our work was accepted as a Spotlight paper to ICML Conference !!! I'm looking forward to presenting our work there this summer and CogSci Society! Big thank you again to my collaborators Wilka Carvalho Yancheng Liang Simon Shaolei Du Max Kleiman-Weiner Natasha Jaques

Famous LLM researcher Bruce Lee quote: "I fear not the LLM who has practiced 10,000 questions once, but I fear the LLM who has practiced one question 10,000 times."

Even with the same vision encoder, generative VLMs (LLaVA) can extract more information than CLIP. Why? Check out our #ACL2025NLP paper led by Siting Li : arxiv.org/pdf/2411.05195

Our new paper tries to uncover what we really need in applying RLVR.

PPO vs. DPO? 🤔 Our new paper proves that it depends on whether your models can represent the optimal policy and/or reward. Paper: arxiv.org/abs/2505.19770 Led by Ruizhe Shi Minhak Song

Congratulations to University of Washington #UWAllen Ph.D. grads Ashish Sharma & Sewon Min, Association for Computing Machinery Doctoral Dissertation Award honorees! Sharma won for #AI tools for mental health; Min received honorable mention for efficient, flexible language models. #ThisIsUW news.cs.washington.edu/2025/06/04/all…

Oral ICML Conference !!! Can't wait to share our work and hear the community's thoughts on it, should be a fun talk! Can't thank my collaborators enough: Wilka Carvalho Yancheng Liang Simon Shaolei Du Max Kleiman-Weiner Natasha Jaques

Check out our new work using online multi-agent RL for LM safety.

I'll present StoryEval tomorrow at CVPR, happy to catch up with new and old friends! 📍ExHall D, Poster #284 ⌚10.30am - 12.30 pm at 6.14

🚨 Code is live! Check out LoRe – a modular, lightweight codebase for personalized reward modeling from user preferences. 📦 Few-shot personalization 📊 Benchmarks: TLDR, PRISM, PersonalLLM 👉 github.com/facebookresear… Huge thanks to AI at Meta for open-sourcing this research 🙌

1/6 Current AI agent training methods fail to capture diverse behaviors needed for human-AI cooperation. GOAT (Generative Online Adversarial Training) uses online adversarial training to explore a pre-trained generative model's latent space to generate realistic yet challenging