Mask Your Logs, Not Your Prompts
· 7 min read
Privacy · AI News
Mask your logs, not your prompts
Redact before the modelRedact before the logs
Almost every "secure your LLM app" guide gives the same advice: before a prompt reaches
the model, strip the personal data out of it — swap names, emails, and account numbers for
[REDACTED] or <PERSON>, then call the model.
It sounds obviously right. I assumed it was, too. Then I went and read the research on what masking actually does to a model, and the papers point the other way. The short version: mask your logs, not your prompts.