Ein LLM-basierter Chatbot als Baustein im Inverted Classroom

This article looks at a chatbot based on Large Language Models (LLM)
for the course "Web Technologies". The chatbot is integrated into the "Inverted Classroom"
concept of the course and made available to students during the current semester.
to the students during the current semester. LLMs with Retrieval Augmented Generation (RAG) offer great potential for the adaptive
preparation and provision of teaching and learning materials. These processes are being evaluated,
the acceptance of the students towards the chatbot, the correctness of the answers and the
the exercises generated by the chatbot in terms of their suitability for the learning objectives,
clarity and difficulty. Overall, it can be seen that students use the chatbot
primarily for knowledge questions and the generation of new exercises. The chatbot was able to
good results with around 80% correct answers and was positively received by the students.
positively received by the students. Overall, the tasks generated were considered to be of appropriate difficulty, clearly
clearly and unambiguously formulated and appropriate to the learning objectives. The insights gained
suggest that task generation should be improved in further work and that the chatbot should be used in other
courses in order to cover a wider range of users.

ChatbotEvaluationArtificial IntelligenceLarge Language ModelsRetrieval- Augmented Generation