Prompt engineering

Principle 1: Clear and Specific Instructions  

Giving clear and precise instructions helps the AI understand exactly what is expected, leading to more accurate and relevant responses.

Tactic 1: Use Delimiters

Delimiters such as quotation marks, triple backticks, or brackets are used to clearly separate instructions from other text. This reduces confusion and improves clarity.

Tactic 2: Structured Output

Requesting responses in a specific format (such as bullet points, tables, or numbered lists) ensures the output is organized, readable, and easy to use.

Tactic 3: Conditional Output

Conditional instructions tell the AI how to respond in different situations. For example, asking it to reply differently based on whether certain conditions are met.

Tactic 4: Few-Shot Prompting

Few-shot prompting involves providing examples within the prompt so the AI can understand the expected style, format, or type of response.

Principle 2: Give the Model Time to Think

Allowing the AI more time to reason helps it produce deeper, more accurate, and well-thought-out answers.

Tactic 1: Ask for Step-by-Step Reasoning

Encouraging the model to explain its reasoning step by step helps break down complex problems and improves accuracy.

Tactic 2: Let the Model Work It Out Internally

Instructing the model to think through the problem before answering leads to clearer logic and better final responses.

What makes a good prompt!!