Inference Parameters Quiz

5 questions Pass: 70% +25 pts (50 with Pro)

Quiz covering Foundation Model Design

What you'll be asked

  1. Which parameter controls the randomness of the model's output by flattening or sharpening the probability distribution of the next token?
  2. You notice that your model is frequently repeating the same phrases in long-form generation. Which parameter should you adjust to discourage this behavior?
  3. When using Top-P (Nucleus) sampling with a value of 0.9, how does the model select the next token?
  4. A user wants a model to be strictly factual and avoid creative hallucinations. Which combination of settings is most appropriate?
  5. If you set 'Top-K' to 50 and 'Top-P' to 0.9 simultaneously, how does the inference engine typically apply these constraints?

Answers and explanations come after you submit.