LLM Lecture: A Deep Dive into Transformers, Prompts, and Human Feedback
The first 500 people to use my link will receive a one month free trial of Skillshare! Get started today! 👉 https://skl.sh/aicoffeebreakwithletitia01251 📺📺This video is an educational and historical deep dive into LLM research, where we would like to show you and convince you that while ChatGPT seemingly emerged overnight, it’s built on years of foundational work and innovation on LLM technology. AI Coffee Break Merch! 🛍️ https://ai-coffee-break-with-letitia.myspreadshop.de/ Thanks to our Patrons who support us in Tier 2, 3, 4: 🙏 Dres. Trost GbR, Siltax, Vignesh Valliappan, Michael, Sunny Dhiana, Andy Ma Outline: 00:00 Lecture contents 01:30 Skillshare (Sponsor) 02:54 The Transformer 05:49 Tokenization 10:11 The Transformer layer 14:04 Attention 22:31 Position embeddings 27:42 Residual connections 29:55 How transformers learn language 34:40 Training Decoders (e.g., GPT) 40:46 Decoder inference / test time 43:55 Encoders (e.g., BERT) 47:45 Encoder-Decoders (e.g., T5). 53:24 Why LLMs need prompting 57:00 Vanilla prompting 01:01:35 Prompt tuning 01:03:03 In context (few-shot) learning 01:07:50 Chain-of-Thought 01:10:25 Retrieval Augmented Generation (RAG). 01:12:14 Beyond pre-training: Post-training techniques 01:13:50 Instruction Tuning 01:16:00 Preference Tuning with Human Feedback via DPO 01:19:20 RLHF key idea 01:21:14 Benchmarking LLMs: Emergent capabilities and how (not) to measure them 01:24:50 Multimodal extensions of LLMs 01:28:04 Gopnik’s Parable of Stone Soup AI 📑Allison Gopnik's Stone Soup AI: https://simons.berkeley.edu/news/stone-soup-ai 📑Alexei Efros’s talk at the @LaureateForum : https://youtu.be/K0gavpe0uDo?si=FdsURxkV39fyU1NV ▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀ 🔥 Optionally, pay us a coffee to help with our Coffee Bean production! ☕ Patreon: https://www.patreon.com/AICoffeeBreak Ko-fi: https://ko-fi.com/aicoffeebreak Join this channel as a Bean Member to get access to perks: https://www.youtube.com/channel/UCobqgqE4i5Kf7wrxRxhToQA/join ▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀ 🔗 Links: AICoffeeBreakQuiz: https://www.youtube.com/c/AICoffeeBreak/community Twitter / X: https://twitter.com/AICoffeeBreak LinkedIn: https://www.linkedin.com/in/letitia-parcalabescu/ Threads: https://www.threads.net/@ai.coffee.break Bluesky: https://bsky.app/profile/aicoffeebreak.bsky.social Reddit: https://www.reddit.com/r/AICoffeeBreak/ YouTube: https://www.youtube.com/AICoffeeBreak Substack: https://aicoffeebreakwl.substack.com/ Web: https://explanationmark.de/letitia https://aicoffeebreak.com #AICoffeeBreak #MsCoffeeBean #MachineLearning #AI #research Video editing: Nils Trost