Why LLMs Need Reinforcement Learning (The Alignment Problem)

Reward Models Explained: Turning Preference into a Number - RL for LLMs

RLHF & InstructGPT: The Recipe Behind ChatGPT RL for LLMs

Policy Gradients & REINFORCE, from Scratch Rl for LLMs

PPO Explained: The Clip and the KL Leash RL for LLMs

RLAIF & Constitutional AI: When AI Grades AI RL for LLMs

DPO: Alignment Without Reinforcement Learning

Rejection Sampling & Best-of-N: The Simplest RL. Rl for LLMs

Process Reward Models: Grading Every Step RL for LLMs

GRPO: Reasoning RL Without a Critic

o1 & Test-Time Compute: Teaching Models to Think RL for LLMs

DeepSeek-R1 & RLVR: Reasoning from Pure RL Rl for LLMs

Agentic RL: Rewarding Multi-Step Tool Use RL for LLMs

RLHF vs DPO vs GRPO: Which Method, and When RL for LLMs