Course

Conceptual LLM Mechanics

Understand how language models work: from training to tokens to inference.

0/7 modules complete0%

Modules

7 total

Understand how pretraining, post-training, fixed weights, and inference-time context shape assistant behavior.

Understand that LLMs process token sequences and generate one token at a time.

Build a conceptual model of transformers, attention, and context influence.

Explain why prompts, examples, and retrieved context influence behavior.

Connect temperature, sampling, and output variability to evals.

Connect prefill, decode, KV cache, streaming, output length, latency, and cost.

Build a support-ticket router and practice deterministic boundaries around probabilistic model output.