I understand, that trace_include_sensitive_data allows to NOT trace agent activity if configured accordingly.
But if mistakenly sensitive data was traced/recorded, seems there is no ability via UI or API to select the affected tracing data and DELETE it - either manually or programmatically.
This should be amended, as it is a typical situation that in error data is recorded that must be purged upon request or observation.
Thank you @OpenAI_Support
Kind regards