Evaluating the Ability of Large Language Models to Generate Motivational Feedback

The paper describes and evaluates the use of large language models (LLMs) to provide personalized motivational feedback in the context of Intelligent Tutoring Systems (ITS). Specifically, the main contributions of the present work are the definition of a novel evaluation framework and the early application of such a framework to assess the ability of LLMs to generate textual feedback including motivational features. The experimentation results show that LLMs demonstrate a promising ability to generate motivational feedback and, therefore, a good chance to be integrated as an additional model into the traditional ITS architecture.