Comment by sullivanmatt

If I were a user or integrator, how do I know that the de-identification step is actually working? Is there a way to test (and/or continue testing) your regex patterns or whatever mechanism used continues to accurately strip my sensitive information before it goes to OpenAI?