✨ New rule - Do you use the system prompt?

categories/artificial-intelligence/rules-to-better-gpt.md

@@ -18,4 +18,5 @@ index:
 - website-chatbot
 - use-semantic-kernel
 - use-embeddings
+- use-system-prompt
 ---

rules/use-system-prompt/rule.md

@@ -0,0 +1,40 @@
+---
+type: rule
+title: GPT - Do you use the system prompt?
+uri: use-system-prompt
+authors:
+  - title: Jack Reimers
+    url: https://www.ssw.com.au/people/jack-reimers
+created: 2023-09-28T12:00:00.000Z
+guid: d98b6b99-28a2-4be3-8095-6f5c45094a98
+---
+
+When you're building a custom AI application using a GPT API you'll probably want the model to respond in a way that fits your application or company.
+You can achieve this using the system prompt.
+
+<!--endintro-->
+
+### What is the system prompt?
+
+Requests to and from a GPT API generally have three types of messages, also known as roles or prompts.
+
+**User**  
+User messages are any messages that your application has sent to the model.
+
+**Assistant**  
+Assistant messages are any messages that the model has sent back to your application.
+
+**System**  
+The system prompt is sent with every request to the API and instructs the model how it should respond to each request.
+When we don't set a system prompt the user can tell the model to act however they would like it to:
+
+![Figure: GPT's responses without a system prompt](without-system-prompt.png)
+
+![Figure: Responses with a system prompt](with-system-prompt.png)
+
+::: greybox
+**Note**  
+Depending on the model you're using, you may need to be more firm with your system prompt for the model to listen. Test your prompt using [OpenAI's Playground](https://platform.openai.com/playground) before deploying.
+:::
+
+For more information on system prompts, see [OpenAI's documentation](https://platform.openai.com/docs/guides/gpt-best-practices/strategy-write-clear-instructions), or use their [playground](https://platform.openai.com/playground) to start testing your own!