What Is ChatGPT Doing … and Why Does It Work?—Stephen Wolfram Writings
Stephen Wolfram explores the broader picture of what's going on inside ChatGPT and why it produces meaningful text. Discusses models, training neural nets, embeddings, tokens, transformers, language syntax.
Contents Top It's Just Adding One Word at a Time Where Do the Probabilities Come From? What Is a Model? Models for Human-Like Tasks Neural Nets Machine Learning, and the Training of Neural Nets The Practice and Lore of Neural Net Training "Surely a Network That's Big Enough Can Do Anything!" The Concept of Embeddings Inside ChatGPT The Training of ChatGPT Beyond Basic Training What Really Lets ChatGPT Work? Meaning Space and Semantic Laws of Motion Semantic Grammar and the Power of Computational Language So ... What Is ChatGPT Doing, and Why Does It Work? Thanks Additional Resources What Is Ch
saved by
- Yung-Hsuan Yang
- Michelle Pan
- Ratan Kaliani
- Asma Lamgh
- Shivam Siddaiya
- Aryan Naik
- Adithya V
- Matviy Kotolyk
- jeremy h
- Nikki Guo
- Saeejith Nair
- Lundeen Cahilly
related reading
- What Is ChatGPT Doing … and Why Does It Work?-Stephen Wolfram Writingswritings.stephenwolfram.com
- Can AI Solve Science?-Stephen Wolfram Writingswritings.stephenwolfram.com
- Deep Dive into LLMs like ChatGPTyoutube.com
- RLHF: Reinforcement Learning from Human Feedbackhuyenchip.com
- What If We Had Bigger Brains? Imagining Minds beyond Ours-Stephen Wolfram Writingswritings.stephenwolfram.com
- The Unreasonable Effectiveness of Recurrent Neural Networkskarpathy.github.io
- Let's build GPT: from scratch, in code, spelled out.youtube.com
- Think of language models like ChatGPT as a “calculator for words”simonwillison.net
- Wolfram|Alpha as the Way to Bring Computational Knowledge Superpowers to ChatGPT-Stephen Wolfram Writingswritings.stephenwolfram.com
- GPT in 60 Lines of NumPy | Jay Modyjaykmody.com
- How LLMs Work, Explained Without Math - miguelgrinberg.comblog.miguelgrinberg.com
- Extracting Training Data from ChatGPTnot-just-memorization.github.io