Three open educational datasets for researchers, developers, and AI tutor builders. All data licensed under CC-BY-4.0.
| Name | Default | Values |
|---|---|---|
| dataset | all | misconceptions | learning-graph | standards | interop | all |
| domain | all | physics | math | biology | chemistry | cs (misconceptions only) |
| format | json | json | jsonl (misconceptions only) |
K-12 STEM misconception ontology. 423 constructs across 9 domains (physics, biology, chemistry, math, cs, earth science, ELA, health, history). Each entry includes the student belief, scientific reality, counterexample, diagnostic question, trigger patterns, severity, persistence, research source, and teacher intervention.
Skill adjacency graph connecting simulation modes. 1,884 directed edges encoding prerequisite relationships, shared concepts, and misconception overlap between modes. Each edge includes priority level (high/medium/low), concepts, misconception references, and student- and teacher-facing explanations.
NGSS standards mapped to simulation modes. Each entry links a mode to a standards code, description, grade band, disciplinary core idea, and discipline.
Crosswalk mapping 118 math misconceptions to ASSISTments practice assignments and 31 misconceptions to Learning Commons Knowledge Graph standard codes. Each ASSISTments link is a deep URL to free remediation practice. Each Learning Commons link maps to a CCSS or NGSS standard code usable with the Knowledge Graph's find_misconceptions_for_standard() tool.
qlm-measure — the public interface layer for the open measurement infrastructure. Evidence events, record verification, and dataset clients. Apache-2.0.
npm install @qlm/measurepip install qlm-measureBuilding the open measurement layer for education AI.
All data licensed under CC-BY-4.0. SDK licensed under Apache-2.0.
For research collaboration: hello@quantumlearningmachines.com