Instruction Tuning Quiz

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

Quiz covering Fine-Tuning and Customization

What you'll be asked

  1. What is the primary goal of instruction tuning in large language models?
  2. When performing instruction tuning, what is the most significant risk associated with using a very small, biased dataset?
  3. Which of the following describes the 'PEFT' (Parameter-Efficient Fine-Tuning) approach in the context of instruction tuning?
  4. Why is it common practice to format instruction-tuning data using specific templates (e.g., '### Instruction:', '### Response:')?
  5. If you are fine-tuning a model on a domain-specific dataset (e.g., legal documents) and notice the model starts performing well on legal tasks but loses its ability to summarize general text, what phenomenon are you observing?

Answers and explanations come after you submit.