Thomas F. Officer
The "system message" provides the LLM with instructions on how to behave.
OpenAI's documentation provides best practices for using its LLMs, a number of which relate to system messages. Many others have shared their experience using system messages. The purpose of this blog post is to collect, clean, and organize all this advice.
The following is a list of system message snippets I've stumbled upon online.
My experience with these 'setting expectations' instructions is that the model is more verbose. It provides a more thorough, detailed response that includes context and reasoning before answering the primary question. Note: depending on your context, this may be more annoying than useful.
Parting thought: I haven't tried this yet, but OpenAI's documentation explains how you can use system instructions to help the model classify the user's request by applying one of a hard-coded sequence of options. Then, you could use that classification to replace the system instructions with something tailored to the type of question the user is asking. See Strategy: Split complex tasks into simpler subtasks.