Quiz me — The Inference Forward Pass

Q1.Which definition of "Prefill vs. decode: two very different phases" matches what the module established?

Q2.For "Prefill vs. decode: two very different phases", which detail or constraint from the module is accurate?

Q3.For "Your first local inference run", which detail or constraint from the module is accurate?

Q4.Which statement about how "Where the FLOPs and the time actually go" actually works is correct?

Q5."Stop conditions and max tokens" — which of these claims is supported by the module?

Q6.Which statement about how "What happens when you call an LLM" actually works is correct?

Q7.When applying "Logprobs at serve time" in practice, which of these holds?

Q8.When applying "From logits to tokens: sampling" in practice, which of these holds?

Q9.Which fact about "Tokenization at serve time" matches the mechanism the module covered?

Q10.Which of these correctly identifies the role of "What happens when you call an LLM" in the broader system?