Support for OpenGPT-X Models configuration?

[Kordishal]

I am trying to use the Chat UI interface to communicate with a vLLM server that serves the https://huggingface.co/openGPT-X/Teuken-7B-instruct-commercial-v0.4.

I run into the issue that the Jinjia Template for the model does not accept the system prompt. It is possible to disable the role with "systemRoleSupported": false,. That however just changes the prompt to use the User role instead. This is not accepted by the model because it always expects alternating User / Assistant role messages. When the system prompt is added as user then there is user message twice in a row. Is there a way to fully remove the system prompt? making the field empty just adds the empty content message instead of none. Is there something I am missing or should I make a feature request?

Or I suppose alternatively if someone can point me in the direction of being able to change the Jinjia template of the Teuken model?