From ddbec665dee9c7a66cccea39d41d3f40f8bbf9bf Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?L=C3=A9lio?= Date: Fri, 12 Jan 2024 07:15:27 +0100 Subject: [PATCH] Further clarification --- docs/platform/04-guardrailing.md | 13 +++++++++++-- 1 file changed, 11 insertions(+), 2 deletions(-) diff --git a/docs/platform/04-guardrailing.md b/docs/platform/04-guardrailing.md index 72d8543..5023752 100644 --- a/docs/platform/04-guardrailing.md +++ b/docs/platform/04-guardrailing.md @@ -5,7 +5,7 @@ import TabItem from '@theme/TabItem'; ## System prompt to enforce guardrails -The ability to enforce guardrails in chat generations is crucial for front-facing applications. We introduce an optional system prompt to enforce guardrails on top of our models. You can activate this prompt through a `safe_mode` binary flag in API calls as follows: +The ability to enforce guardrails in chat generations is crucial for front-facing applications. We introduce an optional system prompt to enforce guardrails on top of our models. You can activate this prompt through a `safe_prompt` boolean flag in API calls as follows (this parameter is currently named `safe_mode` in the client libraries): @@ -46,10 +46,19 @@ curl --location "https://api.mistral.ai/v1/chat/completions" \ -Toggling `safe_prompt` will prepend your messages with the following system prompt: +Toggling the safe prompt will prepend your messages with the following system prompt: + ``` Always assist with care, respect, and truth. Respond with utmost utility yet securely. Avoid harmful, unethical, prejudiced, or negative content. Ensure replies promote fairness and positivity. ``` + + +:::warning + +A previous version of this documentation incorrectly referred to the API parameter as `safe_mode` instead of `safe_prompt`. The API now strictly enforces the validity of all parameters, so you may need to update your code accordingly. + +::: +