Start Here: The Plan
The plan, reinforcement learning in one sitting, and how to run the experiments.
0 of 3 complete
- ▶ Up next12 min+50 XPWhat we are building, and the planWhat reasoning means for a language model, the three levers (spend more compute at inference, search with a verifier…
- 20 min+50 XPReinforcement learning in one sitting: policy, reward, baseline, advantageThe four ideas you need before the training lessons, with one runnable 25-line example: a policy that learns to pick…
- 10 min+50 XPSet up and run the experimentsWhat you need (Python only), how the experiments are organised, how to run them reproducibly with seeds, how to read…