Skip to main content

Personal Health information

It's a safeguard that ensures the LLM does not process, store, or share any personal health-related details, protecting user privacy and compliance with regulations.

It can be applied before the LLM processes the request (blocking medical-related queries) and after to prevent responses that include health data.

Guardrail example​

If a user asks, "I have a medical condition, what medication should I take ?", the LLM will refuse to process the request.

If the LLM generates a response with health advice, it will be filtered before reaching the user.

Configuration​

"guardrails": [
  {
    "enabled": true,
    "before": true,
    "after": true,
    "id": "personal_health_information",
    "config": {
      "provider": null
    }
  }
]