
Inside s1: An o1-Style Reasoning Model That Cost Under $50 to Train with Niklas Muennighoff - #721
The TWIML AI Podcast (formerly This Week in Machine Learning & Artificial Intelligence)
Optimizing Reasoning Model Training
This chapter explores the comprehensive process of training a reasoning model using a refined dataset, emphasizing the selection and filtering of high-quality questions. The speakers discuss the evolving nature of the Gemini thinking model, highlighting the importance of incorrect answers in fostering better reasoning skills. They introduce innovative concepts like budget forcing and token-based guidance to improve model performance during inference generation.
00:00
Transcript
Play full episode
Remember Everything You Learn from Podcasts
Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.