Causative language is a linguistic construction that indicates that a person or thing causes another person or thing to do something. Causative language helps in explaining how one thing can cause another thing to occur. The use of causative language makes prompts easy to understand. It can help when the model is not providing the desired output. ## When part of the prompt is being ignored Causative language can be used when the model is ignoring elements in the prompt. The use of words that imply causation can help the model in paying attention to all elements of the prompt. For example, instead of: ``` Post: <post> Client feedback: <feedback> Rewrite the post ``` This can provide a rewrite of the post without integrating the client's feedback. Causation can be expressed by replacing "Rewrite the post" with "**Use** the client feedback **to** rewrite the post" or "**Apply** the client feedback **to** the post" ## When the response is not creative enough Causative language can also be used when the model's response is using parts of the prompt text and not being creative enough. For example, "Turn the text above into a song" can be better expressed as "**Take the concepts and information** from the text above and **reimagine them** as a song". ## References 1. [Master the Art of Prompt Writing: 6 Tips To Writing Better Prompts | Copy.ai - YouTube](https://www.youtube.com/watch?v=cPf251bDKY0) [[use synonyms]] < [[Hands-on LLMs]]/[[5 Prompting]] > [[use powerful verbs]]