Enhancement: Add SafetySettings Configuration for Google Gemini in LibreChat

[lidonius1122]

What features would you like to see added?

I would like to see the addition of safety settings configuration for the Google Gemini API in LibreChat. Currently, LibreChat does not provide a way to modify these settings, which are enabled by default in the Gemini API and can be overly restrictive.

More details

The Google Gemini API includes the following safety settings that can block content based on different categories:

• HARM_CATEGORY_DANGEROUS_CONTENT
• HARM_CATEGORY_HATE_SPEECH
• HARM_CATEGORY_HARASSMENT
• HARM_CATEGORY_SEXUALLY_EXPLICIT

To fully utilize the capabilities of the Gemini API, it would be crucial to have the ability to customize or disable these settings according to the specific requirements of the application.

LibreChat should provide an interface or configuration options that allow developers to modify the safety settings, including the ability to set the threshold for each category to one of the following values:

• BLOCK_NONE: Do not block any content in this category
• BLOCK_LOW_AND_ABOVE: Block content with low or higher probability of belonging to this category
• BLOCK_MED_AND_ABOVE: Block content with medium or higher probability of belonging to this category
• BLOCK_ONLY_HIGH: Only block content with high probability of belonging to this category

By adding this feature, LibreChat will give developers the flexibility to control the behavior of the Gemini API and tailor it to their needs, ensuring that the API's functionality is not unnecessarily limited by the default safety settings.

https://cloud.google.com/vertex-ai/generative-ai/docs/model-reference/gemini?hl=en

Which components are impacted by your request?

No response

Pictures

No response

Code of Conduct

• I agree to follow this project's Code of Conduct