Being able to create models with personalized prompts.

[photonn]

There is way to add a default prompt for all the models using chat environment variables. We would like to have a more granular approach in which that is changed (or have both options) to a model by model setup. For example, being able to define multiple models of the same type with a "system_prompt" filed in the core config that allow us to create "departament personalized propmpts" without the need to mimic how the chat works with a separate app for each department. for example:

"models": {
    "gpt-35-turbo": {
        "type": "chat",
        "displayName": "HR",
        "overrideName":"hr",
        "overridePrompt":"You are an HR specialist with years of experience selecting the best engineers for your company. Your language is concise and to the point"
        "displayVersion": "Latest",