The same model gives different answers depending on how you ask. Prompt engineering designs inputs to guide a language model toward desired output, using instructions, examples, formatting cues, and role descriptions.

Simple techniques include asking for step-by-step reasoning or providing a few examples before the question. More elaborate methods chain prompts together, ask the model to critique its own output, or decompose a task into smaller parts.

Common prompting techniques

• Zero-shot instructions
• Few-shot examples
• Chain-of-thought reasoning
• Role and persona assignment
• Output format specification

Prompting is brittle. A wording that works on one model may fail on another, and small changes can shift results unpredictably. Treating prompts as code, with version control and evaluation, reduces the chaos.