The Sequence Chat: The Transition that Changes Everything. From Pretraining to Post-Training in Foundation Models
Was this email forwarded to you? Sign up here The Sequence Chat: The Transition that Changes Everything. From Pretraining to Post-Training in Foundation ModelsOne of the most impactful transitions in the generative AI space
The release of GPT-01 marked many important milestones in the generative AI space. The model has sparked a tremendous new phase of innovation in reasoning models which has materialized in the release of models such as DeepSeek’s R1 or Alibaba’s QwQ. The magical reasoning capabilities of these models is powered by an increasing transition from pretraining to post-training computation time. In this essay, we will explore the fundamentals behind that transition highlighting the limitations associated with scaling pretraining and the emerging techniques in post-training. Furthermore, it emphasizes the shift away from traditional reinforcement learning with human feedback (RLHF) towards innovative methodologies that promise to enhance model performance and adaptability. Understanding Pretraining in Foundation Models...Subscribe to TheSequence to unlock the rest.Become a paying subscriber of TheSequence to get access to this post and other subscriber-only content. A subscription gets you:
|
Older messages
Edge 454: Meet Magenctic-One, Microsoft's New Framework for Building Multi Agent Systems
Tuesday, December 10, 2024
Built on AutoGen, the framework is designed for agents that collaborate in open ended tasks. ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏
World Models are Coming and They are Awesome
Tuesday, December 10, 2024
Two amazing world models were released this week. ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏
📝 Guest Post: Advanced RAG Techniques: Bridging Text and Visuals for More Accurate Responses*
Tuesday, December 10, 2024
In this guest post, Fendy Feng from ZIlliz explores how RAG works, RAG challenges, and advanced RAG techniques like Small to Slide RAG and ColPali. ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏
Edge 453: Distillation Across Different Modalities
Tuesday, December 3, 2024
Cross modal distillation is one of the most interesting distillation methods of the new generation. ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏
Alibaba QwQ Really Impresses at GPT-o1 Levels
Sunday, December 1, 2024
The new model matches and surpasses GPT-o1 on reasoning tasks. ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏
You Might Also Like
Reach More Readers, newsletterest1 – BOOST Your Story on HackerNoon🔥
Wednesday, December 11, 2024
Get Your Story Featured on the Homepage and in The HackerNoon Newsletter ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏
Top Tech Deals 👀 $109 Robot Vacuum, Google Pixel Sale, Anker Power Bank, and More
Wednesday, December 11, 2024
Grab a new Pixel phone or tablet, stocking stuffers, and other goodies. How-To Geek Logo December 11, 2024 Top Tech Deals: $109 Robot Vacuum, Google Pixel Sale, Anker Power Bank, and More Grab a new
Hurry, newsletterest1! Less Than a Week Left to Compete for $2,500 in the AI Writing Contest 🏃
Wednesday, December 11, 2024
Start drafting your entry today! ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏
DePIN On Ethereum: Redefining Coordination Systems
Wednesday, December 11, 2024
Top Tech Content sent at Noon! Boost Your Article on HackerNoon for $159.99! Read this email in your browser How are you, @newsletterest1? 🪐 What's happening in tech today, December 11, 2024? The
Post from Syncfusion Blogs on 12/11/2024
Wednesday, December 11, 2024
New blogs from Syncfusion Building a Neumorphic UI with .NET MAUI Column Chart to Showcase Gen Z's Favourite Social Media Platforms By Dhanaraj Rajendran Learn to create a Neumorphic UI with
24 Hours Until Our 2025 Outlook Webinar – Register Now ⏰
Wednesday, December 11, 2024
Don't miss the key trends shaping 2025 with our free webinar on December 12th. View Online | Subscribe | Download Our App FREE WEBINAR - Tomorrow at 11am PST 2025 Outlook: Key Trends on Our Radar
⚙️ Another AI lawsuit
Wednesday, December 11, 2024
Plus: Tesla sued ... again
The most Windows-like Linux distro
Wednesday, December 11, 2024
iOS 18.2 arrives; AI moves undercover; Natural Cycles dupe -- ZDNET ZDNET Tech Today - US December 11, 2024 The default Wubuntu desktop. This Linux distro is so Windows-like, it even comes with
Your InfoSec Survival Guide
Wednesday, December 11, 2024
How to optimize your compliance practices through a continuous monitoring approach The Hacker News The InfoSec Survival Guide Today, security and compliance leaders are struggling under the pressure of
The Sequence Chat: The One Area in Which China can Dominate the US in the AI Race
Wednesday, December 11, 2024
Might come as a surprise. ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏