Is it possible to give one predicted class of model (RF) to LLM and generate descriptions about that class? For example, if RF predicts a class as dirty water, then LLM must be able to tell the user the drawbacks of dirty water or something like that. In short, I am trying to integrate RF with LLM.
Yeah i think that is possible, just feed the prompt for a description to the LLM and of the LLM has been trained on that area it will give you a description of the class.