Reading
Articles I’ve read and think are worth your time.
2026
- Reinforcement Learning for LLMs: The Complete GuideThe most thorough intro to RL for LLMs I’ve read. It starts from value functions, advantages, KL and importance sampling, works through REINFORCE / RLOO and PPO, then GRPO and its follow-ups (DAPO, GSPO, Dr. GRPO, CISPO), and ends on rubric rewards, RL scaling laws and agentic RL. Full derivations and PyTorch code throughout. It’s long, so treat it as a reference to come back to.