Transformers & Attention
Ten questions that probe the mechanics behind self-attention, positional encoding, KV caching, and the design choices that make transformers work in practice.
Test your ML depth. Each quiz covers a focused topic with 10 questions — answer, see the explanation, and track your score at the end.
Ten questions that probe the mechanics behind self-attention, positional encoding, KV caching, and the design choices that make transformers work in practice.
Ten questions on the mechanics of gradient descent, adaptive optimizers, learning rate schedules, gradient pathologies, and the numerical tricks that keep large model training stable.
Ten questions on the policy gradient theorem, PPO, reward modelling, RLHF, KL penalties, and the RL concepts that underpin modern LLM alignment and fine-tuning.
Ten questions on the mathematical foundations underlying modern ML — probability, information theory, linear algebra, and the statistical reasoning behind common model design choices.