Use LLM to Assess Feedback Quality

I’m working on a project to assess the quality of feedback given on a forum. Basically, someone posts a proposal or plan on the forum, and others can ask questions, and give suggestions and criticisms.

Is there a way to use LLM to quantify the overall helpfulness of all the feedback / entire discussion? Would the right approach be to provide the LLM with a rubric that grades feedback from, say, 1-5, with a description of what each of the numbers mean?

Hi @Guangye_Cao,
Welcome to the community. I’ve moved your post to AI questions to allow the right people to find and provide answers to your questions.