🔍 Ever wondered how ChatGPT actually works? In this video, we break down the architecture and training process behind ChatGPT, including Supervised Learning, Reinforcement Learning, and RLHF (Reinforcement Learning from Human Feedback).
_____________
*Chapters:*
0:00 - Introduction to training methods
0:27 - Capability vs Alignment
01:01 - Examples of Misaligned model
01:27 - Common Misalignment Issues
01:58 - Training Strategies leading to misalignment
02:04 - Next Token Prediction
02:30 - Masked Language Modeling (MLM)
02:54 - MLM problems
03:30 - How ChatGPT solved Misalignment Problem?
03:41 - Supervised Fine Tuning
04:20 - Reward Model
05:11 - What is PPO?
06:19 - Performance Evaluation
06:49 - Outro

_____________
Other tutorials:
_____________

*Build custom Chrome Extension to Improve Grammar using AI:* https://youtu.be/9REhki2hhGg

*How to install python & IntelliJ:* https://youtu.be/t2MVd8R2hGs

*How to use LLM in local machine:* https://youtu.be/SroglXNjHgc

*Build first AI Assistant using LangChain:* https://youtu.be/uPlBKAQAcCE

*Build AI ChatBot using Streamlit:* https://youtu.be/IPOv0hqJ-wQ

----------------------