The Video Assistant · Lesson 4 — The Examples ← Course

The examples — few-shot prompting.

Telling the model what you want is good. Showing it a couple of solved examples is better.

Last lesson you described the job in words. There is an even more reliable way to get the answer you want: show the model a few examples of the job already done, then ask. Asking with no examples is called zero-shot; giving a handful first is few-shot.

Picture showing the VAR two earlier calls already settled — one a penalty, one waved away — each written up the same way. “Here is how we call these.” Faced with a fresh incident, the VAR naturally answers in the same shape.

It works because the model is, underneath, a pattern-continuer — the idea from Lesson 1. Show it a pattern and it carries the pattern on. A worked example pins down the format, the level of detail, even the tone far more precisely than a sentence of description can.

Nothing is retained. The examples work only because they are sitting in the prompt — start a new conversation without them and the model is back where it began.

This is the quickest fix when the model nearly does what you want but not in the right shape. Two or three good examples often beat a paragraph of instructions — and they teach by demonstration, the tricky cases included.

Show the VAR zero, then one, then two prior calls — and watch the new verdict snap into the same shape.

No examples yet — zero-shot. The VAR answers in its own words.
prior calls shown first · examples
↓ then the real one ↓
the new incident · to be ruled
↓
the VAR's call

Examples first, then the question.

A few-shot prompt is literally that: the solved examples, then the real incident, handed over together. The model continues the pattern it was shown. 


      

    


    

      
Show, don't just tell.

      

        Describing the format you want is good; demonstrating it is better. A couple of
        solved examples pin down the shape, the detail, and the tone more precisely
        than a sentence can — and the model simply continues the pattern. Add a
        contrasting example and the shape widens — the model has a NO PENALTY and a lower
        confidence in its repertoire, not just one verdict.
      

      

        Reach for few-shot when the model nearly does what you want but not in the
        right shape. Show it two or three clear, varied examples — the tricky calls
        included — and let it copy.
      

      

        Next: how many? →
        
      

    


  


  

    

      ▸ Go deeper — getting few-shot right optional
      

        
What counts as a “shot”

        

          One “shot” is one worked example — a single solved case. Show none and it is
          zero-shot; one is one-shot; a small handful is few-shot. The word just counts
          how many finished examples you place before your real question.
        

        
Where the examples sit

        

          The examples ride in the prompt itself, alongside the system briefing and your
          incident — the same context window from Lesson 2. They are part of the input for
          this one request, not stored anywhere the model keeps.
        

        
Order matters

        

          The model leans hardest on the most recent examples, so put your clearest,
          most representative call last. If two examples disagree, the later one tends
          to win.
        

        
Examples now, or training later

        

          Few-shot teaches a pattern for this one conversation, right in the prompt. If
          you need the model to always know the pattern without being shown each time,
          that is a different tool the course reaches near the end.