Feedback taken. We are working on making it more explicit for users to mention the task and dataset they want to train models on. Additionally, we will introduce a flag to let people mention the prompt they want to use for finetuning these models.