Week 11 — Reasoning Models & RL

Material	Where
Lecture notes	published before class
Lab	published before class
Papers	STaR (Zelikman et al., 2022) (PDF) · DeepSeek-R1 (2025) (PDF)

This week

Reasoning models internalize the caller-side procedures of W2: thinking budgets as a billable resource, the training problem (no process labels), STaR bootstrapping, reinforcement learning fundamentals (policy, reward, RLHF vs. RLVR), and DeepSeek-R1 with GRPO.

Lab plan

Weekly LangChain and LangGraph practice follows LangChain Academy and official RAG tutorials.