Генерація тестових завдань з текстових навчальних матеріалів

Abstract

The development of an automated test generation system using large language models (LLM) such as GPT and Perplexity (Sonar) offers significant advancements in education. This innovative system automates the creation of test questions from various educational materials, generates realistic distractors, and classifies questions by difficulty level. By enhancing assessment objectivity and reducing educator workload, it creates more effective tests while ensuring comprehensive evaluation of students' knowledge and critical thinking skills.