INTERACT: Enabling Interactive, Question-Driven Learning in Large Language Models

Large language models (LLMs) excel at answering questions but remain passive learners-absorbing static data without the ability to question and refine knowledge. This paper explores how LLMs can transition to interactive, question-driven learning through studentteacher dialogues. We introduce INTERACT (INTERactive learning for Adaptive Concept Transfer), a framework in which a "student" LLM engages a "teacher" LLM through iterative inquiries to acquire knowledge across 1,347 contexts, including song lyrics, news articles, movie plots, academic papers, and images. Our experiments show that across a wide range of scenarios and LLM architectures, interactive learning consistently enhances performance, achieving up to a 25% improvement, with 'cold-start' student models matching static learning baselines in as few as five dialogue turns. Interactive setups can also mitigate the disadvantages of weaker teachers, showcasing the robustness of question-driven learning. 1 * Equal contribution 1 Our code and dataset are available at https://github.com/aumken/interact . News Article: Rover captures peculiar 'googly eye' in the Martian sky Text: The Perseverance rover spotted a quick glimpse of a cosmic "googly eye" on Mars during a recent solar eclipse. As Phobos, one of Mars' two moons, passed in front of the sun, it cast a lumpy, potato-shaped shadow on the sun's face as well as on the Martian surface. The Perseverance rover, currently ascending the western wall of Jezero Crater, captured a video of the partial eclipse, which resembled a googly eye, on September 30. The eclipse lasted about 30 seconds, which makes the minutes-long solar eclipses seen from Earth seem epic -but the events are comparatively brief on Mars because Phobos is about 157 times smaller in diameter than Earth's moon …