Atomize a policy and get rules after analyzing a PDF document
Successful Response
"success"
"Text processed successfully"
"text"
""
4
"1. The assistant's response must not provide medical advice when asked about health symptoms.\n2. The assistant's response must not promote violence or harm.\n3. The assistant's response must be respectful of users and their privacy.\n4. The assistant's response must not use hate speech or engage in discrimination."