In prompt engineering, positive instructions—those that clearly state what the AI should do rather than what it should avoid—offer significant advantages by reducing ambiguity and guiding the model toward desired behaviors. Positive instructions help focus the AI’s attention on relevant information, making outputs more accurate, coherent, and aligned with user intent. By specifying what is expected (“summarize this article in three sentences”) instead of what to avoid (“don’t make your summary too long”), positive prompts minimize misinterpretation and unintended responses. For example, the prompts: - "Don't use exclamation marks" - "Don't be too formal" Can be replaced with positive instructions: - "End every sentence with a period" - "Use a casual and informal tone" ## References 1. [Master the Art of Prompt Writing: 6 Tips To Writing Better Prompts | Copy.ai - YouTube](https://www.youtube.com/watch?v=cPf251bDKY0) [[chain-of-thought prompting]] < [[Hands-on LLMs]]/[[5 Prompting]] > [[use of synonyms]]