🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

I am actively seeking industrial opportunities and bring expertise in training T2I and T2V diffusion models, along with a strong background in Deep RL. If you think my skills align with your needs, feel free to reach out!

Michael is a rising star, fun labmate and great person! Please consider hiring him. He will surprise you with his Mandarin skill 🤓

I am deeply sorry and heartbroken over the loss of Felix Hill. His post docs.google.com/document/d/1aE… is a poignant reminder of the mental health challenges we face in the fast-paced and high-pressure AI field. Lately, I’ve also been feeling overwhelmed by the rapid advancements in

This "Aha moment" in the DeepSeek-R1 paper is huge: Pure reinforcement learning (RL) enables an LLM to automatically learn to think and reflect. This challenges the prior belief that replicating OpenAI's o1 reasoning models requires extensive CoT data. It turns out you just

🚀 Excited to share that our work has been accepted to #NAACL2025! We show that LLM watermarks can be removed in a black-box setting 🛠️ For more details: arxiv.org/abs/2411.01222

Let me try it

We are reproducing the full DeepSeek R1 data and training pipeline so everybody can use their recipe. Instead of doing it in secret we can do it together in the open! 🧪 Step 1: replicate the R1-Distill models by distilling a high-quality reasoning corpus from DeepSeek-R1. 🧠

Are there any papers that theoretically or quantitatively demonstrate that training a language understanding model, like a metric or reward model, is easier than training a language generation model? Alternatively, should I justify this based on the differences in the output

A must read paper!

The best knowledge distillation tutorial so far! Also include speculative knowledge distillation : ) .

a newly baked Dr. Congratulations to Wenda Xu for successfully defending his phd thesis "On Evaluation and Efficient Post-training for LLMs". Highly recommend his slides: covering RL training, better KD, LLM/text gen evaluation, bias in LLM as a judge: docs.google.com/presentation/d…

Finally completed this journey. Thanks a lot to my family, advisors and friends!