In the final graded lab of Module 5 we define user prompt for research_agent and system prompt for writer_agent. Is that intentional? Why can’t we have same prompting structure for both agents?
1 Like
Module 5 of what Course?
I checked Prompt Engineering with Llama 2 (the Course section in which you posted your question), and it only has a graded quiz. No graded lab nor multiple modules, it’s a Short Course.
Course is Agentic AI by AndrewNg, Module 5.
+1, I was also wondering the same. Is there a reason why there are both system and user prompts starting from Module 5? The previous modules only specify user prompts and it would be a great learning data point on why we need to specify both for multi-agents.
Check out this post for some thoughts about system vs user prompts: System and user messages: why need both? - #2 by lihanguang
1 Like