StatusKT Pipeline Uses LLMs to Extract Mathematical Proficiency
September 9, 2026
The StatusKT framework uses a three-stage teacher-student-teacher LLM pipeline to transform student problem-solving processes into intermediate mathematical proficiency representations. This method improves Knowledge Tracing by moving beyond binary correctness to capture multidimensional learning states.
HOW THIS AFFECTS YOU
●
builderYou can implement this three-stage pipeline to build more nuanced intelligent tutoring systems.
●
researcherYou can utilize the KT-PSP-25 dataset to train more explainable knowledge tracing models.