SproutStack🌱 AI Engineering · How LLMs Work (Visual First) · cozy lesson
The LLM Training Pipeline 🔒 premium preview
🔒 Marked premium for future. Free while we build locally — payment comes later.
12 min · 1 min read · no scary math, promise
🤖
You’ve got this. Read a little, play a little — I’ll wait. No rush.
Three stages
- Pretrain: next-token on web/books/code. Learns grammar, facts (stale), reasoning patterns.
- SFT: examples of good assistant replies. Learns format.
- Preference (RLHF/DPO): rank answers, push toward helpful/safe.
Why you care: base models complete, chat models follow. Cutoff = pretrain end. Fresh facts need RAG, not bigger pretrain.
💛 Enjoying? Try 5 playful quizzes or watch it move.
Check your understanding
Correct answers earn XP (once each).
1. Pretrain vs SFT?
2. RLHF/DPO for?
My notes (saved in this browser)
Select text above → Save selection, or write your own. AlgoMaster-style notebook, local-first for MVP.
No notes yet. Your highlights will live here.
Finished reading? Seal it with a tick ✅
The checkbox in the explorer turns green too — same progress.