Open-Assistant
Open-Assistant copied to clipboard
Allow per-chat sampling parameters
As I understand, the sampling parameters are currently stored client-side in localStorage and apply to all chats. From a user perspective, this is a bit unintuitive and it would be nice to have control over the parameters on a chat-by-chat basis. E.g. playing around with different parameters to see what gives the best results is very tedious if the settings are global.
Do we have the backend support for GET/POST chat sampling parameters? I don't think using client localStorage would be efficient for this