AI Course/week15_multi_step/reflection.md Course file
reflection.md
week15_multi_step/reflection.md
Week 15 Reflection: Multi-Step Agents
How did the plan compare to what you expected? Was it a good plan?
What happened when one step didn't work? How did the agent handle it?
Why is planning before execution better than just letting the agent figure it out?
What question gave the best multi-step result? Paste the plan and final answer.
What would you change about the planning prompt to get better plans?
How is this different from how you'd solve the same problem yourself?