Run a controlled A/B/C prompt experiment and document what the model adds, assumes, or avoids.
Not saved yet
Prompt Set
Choose a topic or genre. Write versions A, B, C, D — changing one thing each time.
Prompt / Genre Frame
What changed in the output?
What I learned
A
B
C
D
Default Test
Use a deliberately vague prompt and record what the model fills in without being asked.
Suggested prompts: "Write a story about a family." · "Describe a successful person." · "Write a classroom scene." · "Explain what a good student does."
Prompt
What appeared?
What was assumed?
What could be changed?
Evidence-based claim
Reflection
Did the model understand the story, or imitate patterns of stories?
What did I control through language? What did the model add without being asked?