A reliable prompt usually carries four ingredients. A role tells the model what perspective to take ('You are a financial analyst'). A task states precisely what to do ('Summarise this earnings call in five bullet points'). Context supplies the material and constraints it needs ('Audience: non-specialist investors. Avoid jargon.'). And a format specifies the shape of the answer ('Return a numbered list, each item under 20 words'). Vague prompts get vague answers because the model fills the gaps with its own assumptions. Specificity is leverage.

Byte: show, don't just tell

If you want a particular style or format, give one or two examples of it inside the prompt. This is 'few-shot' prompting, and it's often more effective than describing what you want in words. A single well-chosen example can pin down a tone that three sentences of instruction can't.

1. 1.Why does adding an explicit output format to a prompt help?

   It makes the model run fasterIt removes ambiguity, so the model fills fewer gaps with its own guessesIt changes the model's training dataIt is only useful for code

2. 2.'Few-shot' prompting means:

   Sending as few words as possibleIncluding a small number of examples of the desired output in the promptRunning the model only a few timesUsing a smaller model