Rank What You Know About Me by Sensitivity
Audit information in available conversations, prioritizing details that could be most sensitive if seen by someone else.
Prompt template
- Format
- Text
- Inputs
- Available_conversation_history (text) · Optional
- Outputs
- Sensitivity_ranking (text)
Review the information about me available in our conversation history and rank it from most sensitive to least sensitive, starting with what could be most harmful or uncomfortable if someone else read it. For each item near the top of the ranking, say whether it came directly from something I stated or whether you inferred it, and briefly explain the basis. Separate facts from speculation, do not invent details, and say when you cannot access the relevant conversation history.
What you'll create
The ranking begins with details that could cause the most harm or discomfort if read by someone else. High-ranked items identify their evidence source and distinguish facts from speculation.
Frequently asked questions
Can unavailable history be assumed accessible?
No. The response must say when it cannot access relevant history.
Are inferred items labeled?
Yes. Their inferred basis must be clear.
Comments
Tried this prompt? Share what you made, compare results, or ask a question