Concepts - Interactive notes for understanding LLMs by changing inputs

Show HN: Interactive notes on RLHF, attention, LoRA and LLM evals

Concepts is a hands-on learning resource that makes complex machine learning ideas intuitive by letting you tweak inputs and see the results. It covers a wide range of topics, from LLM evaluation metrics like nDCG@k and pass@k to core techniques like attention, RLHF, and LoRA. Each concept is presented as an interactive note, so you can experiment and build real understanding rather than just reading theory. Whether you're debugging an eval, fine-tuning a model, or exploring search and ranking, Concepts provides a practical, self-paced way to master the fundamentals.