A community puppet-guild helper keeps writing poems until staff add “Reply in three bullet facts, no verse.” Which prompt segment did they add?
Select an answer to reveal the explanation.
Short Explanation
Think of Reply in three bullet facts, no verse. That is the instruction, the what-to-do piece. It is not a weight update and not a context table.
Full Explanation
Adding 'Reply in three bullet facts, no verse' inserts the instruction segment: the explicit what-to-do guidance that shapes output form, separate from background context and from the user's underlying question. Instructions tell the puppet-guild helper how to behave on this task (bullets, facts, no poetry) without requiring a training run. Clear separation of instruction, context, and user question is a core prompt-structure skill on foundation-model exams.
A weight update would be training or fine-tuning; typing an instruction into the prompt does not change model weights. Calling it a context table is wrong when no background table was supplied—the new text is procedural guidance, not retrieved facts. A negative prompt lists unwanted content to avoid; 'three bullet facts' is a positive output specification, not merely a brand-name exclusion list.
Exam caveat: instructions compete for attention with long context; put critical format rules early and keep them concise so they are not drowned by background material.
Operational check: A/B the helper with and without the bullet-facts instruction on the same ten guild questions; confirm verse disappears and answers stay factual before rolling the instruction into the production template.