What I Learned Implementing On-Policy Distillation
An experiment distilling Qwen3-8B into a 0.6B student: the mathematics of teacher feedback, the improvement greedy evaluation nearly missed, and what a failed training run revealed.
Training
From RLHF to GRPO: The RL Techniques That Align Language Models
How reinforcement learning transforms raw language models into useful assistants — from PPO’s four-model pipeline to DPO’s elegant shortcut to GRPO’s reasoning revolution, with the math that makes each one work.