From the course: LLMs for Enterprise: Technical Protocols, Considerations, and Data Privacy

Unlock this course with a free trial

Join today to access over 25,300 courses taught by industry experts.

Redacting sensitive data

Redacting sensitive data

From the course: LLMs for Enterprise: Technical Protocols, Considerations, and Data Privacy

Start my 1-month free trial Buy for my team

Redacting sensitive data

“

- One of the capabilities of large language models is to synthesize answers based on text or data. Some of this data might be quite sensitive, so let's walk through two examples on how to pre-process this data to take advantage of these capabilities. Starting with text, you may want to transform it into a different format. For example, you might want to turn some notes into an email about a magic new product. To maintain the secrecy of the name, you can change it to [SECRET_NAME]. And then after generating the email, copy it and manually replace it. This way, you leverage the power of gen AI and maintain data privacy. One of the newer capabilities of LLMs is analyzing data. For example, OpenAI code interpreters and cloud artifacts can be used to analyze spreadsheets, build visuals, and render them right in the browser. One recommendation in this case is to give fake numbers or only upload data column names. For example, if I have a CSV with the columns Name, Date, Location, Phone…

Contents