2.8 — Demo: Prompt Playground 🧪

🟢 In plain words — one same request, four ways to phrase it. Watch how the prompt changes… and how the output changes with it.

One task, four techniques. Compare the prompt sent and the output for the Zero-shot, the Few-shot, the Chain-of-Thought (reasoning revealed step by step) and the Tree-of-Thoughts (branches explored then pruned).

Prompt Playground— one task, four techniques

The taskCustomer message: "This is the 3rd time my order arrives late, this is unacceptable." → Give the urgency (1-5), the tone, and a short reply.

Zero-shotFew-shotChain-of-ThoughtTree-of-Thoughts

Prompt sent

Classify this customer message (urgency, tone) and propose a reply.
Message: "…order late, unacceptable."

Simulated output

The customer is unhappy. Reply: "We're sorry for the delay, we're looking into it."

Fast, but the format is unstable: no numeric urgency, no explicit tone. Fine for a simple task.

⚠️ Educational simulation in your browser (no real AI called). The point: feel how prompt structure changes output quality.