An Open Course on LLMs, Led by Practitioners

Context

• It is a survey course by design, introducing topics rather than walking through code, so Hamel Husain tells students to apply each part to a personal project or the material won't stick.
• Do the instruction-tuning tutorial even if you never intend to fine-tune: it forces you through working with open-weight models, generating synthetic data, managing prompts, and generating predictions.
• The paid Maven cohort's recordings were re-cut for self-service, with chapter summaries, notes, and slides per talk grouped by subject, so you navigate straight to what you need instead of watching in order.

• eval — A repeatable test for AI quality — a set of tasks plus scoring — used the way software teams use test suites, because model output is too variable to judge by eyeballing.
• RAG — Retrieval-augmented generation — fetching relevant documents first and pasting them into the model's context so it answers from your data instead of memory.
• fine-tuning — Taking a trained model and training it a bit more on your own examples so it gets better at one specific job.
• open weights — A model whose trained parameters are published for anyone to download and run — unlike API-only models you can access but never possess.

• articleDo Automated Evals Work?
• articleWhy I Stopped Using nbdev
• articleEvals Skills for Coding Agents
• articleYour AI Product Needs Evals

• postLLM CourseNeo Kim
• articleLLM Evals: Everything You Need to KnowHamel Husain
• articleAI Engineer 2024 Keynote - What We Learned from a Year of LLMsEugene Yan

Checking sign-in…

Loading comments…