---
pretty_name: QVAC Genesis III
dataset_name: GenesisIII
language:
- en
license: cc-by-nc-4.0
size_categories:
- 100M<n<1B
task_categories:
- text-generation
tags:
- synthetic-data
- education
- STEM
- pretraining
- reasoning
- failure-analysis
- option-level-reasoning
configs:
- config_name: full
  default: true
  data_files:
  - split: train
    path:
    - data/astronomy/preview.parquet
    - data/college_biology/preview.parquet
    - data/college_chemistry/preview.parquet
    - data/college_computer_science/preview.parquet
    - data/college_maths/preview.parquet
    - data/college_medicine/preview.parquet
    - data/college_physics/preview.parquet
    - data/conceptual_physics/preview.parquet
    - data/econometrics/preview.parquet
    - data/electronic_science/preview.parquet
    - data/geography/preview.parquet
    - data/high_school_biology/preview.parquet
    - data/high_school_chemistry/preview.parquet
    - data/high_school_computer_science/preview.parquet
    - data/high_school_maths/preview.parquet
    - data/high_school_physics/preview.parquet
    - data/high_school_statistics/preview.parquet
    - data/machine_learning/preview.parquet
    - data/professional_medicine/preview.parquet
    - data/high_school_maths/train-00000-of-00016.parquet
    - data/high_school_computer_science/train-00000-of-00017.parquet
    - data/college_maths/train-00000-of-00026.parquet
    - data/electronic_science/train-00000-of-00019.parquet
    - data/college_medicine/train-00000-of-00040.parquet
    - data/high_school_physics/train-00000-of-00011.parquet
    - data/machine_learning/train-00000-of-00020.parquet
    - data/college_physics/train-00000-of-00018.parquet
    - data/high_school_chemistry/train-00000-of-00018.parquet
    - data/high_school_statistics/train-00000-of-00017.parquet
    - data/geography/train-00000-of-00019.parquet
    - data/college_biology/train-00000-of-00026.parquet
    - data/high_school_biology/train-00000-of-00032.parquet
    - data/college_chemistry/train-00000-of-00019.parquet
    - data/professional_medicine/train-00000-of-00012.parquet
    - data/conceptual_physics/train-00000-of-00012.parquet
    - data/econometrics/train-00000-of-00018.parquet
    - data/astronomy/train-00000-of-00018.parquet
    - data/college_computer_science/train-00000-of-00019.parquet
    - data/high_school_maths/train-00001-of-00016.parquet
    - data/econometrics/train-00001-of-00018.parquet
    - data/college_computer_science/train-00001-of-00019.parquet
    - data/high_school_biology/train-00001-of-00032.parquet
    - data/geography/train-00001-of-00019.parquet
    - data/high_school_statistics/train-00001-of-00017.parquet
    - data/college_biology/train-00001-of-00026.parquet
    - data/high_school_chemistry/train-00001-of-00018.parquet
    - data/college_medicine/train-00001-of-00040.parquet
    - data/professional_medicine/train-00001-of-00012.parquet
    - data/college_chemistry/train-00001-of-00019.parquet
    - data/astronomy/train-00001-of-00018.parquet
    - data/college_maths/train-00001-of-00026.parquet
    - data/electronic_science/train-00001-of-00019.parquet
    - data/high_school_physics/train-00001-of-00011.parquet
    - data/conceptual_physics/train-00001-of-00012.parquet
    - data/high_school_computer_science/train-00001-of-00017.parquet
    - data/college_physics/train-00001-of-00018.parquet
    - data/machine_learning/train-00001-of-00020.parquet
    - data/high_school_maths/train-00002-of-00016.parquet
    - data/high_school_computer_science/train-00002-of-00017.parquet
    - data/professional_medicine/train-00002-of-00012.parquet
    - data/college_medicine/train-00002-of-00040.parquet
    - data/geography/train-00002-of-00019.parquet
    - data/college_chemistry/train-00002-of-00019.parquet
    - data/college_maths/train-00002-of-00026.parquet
    - data/electronic_science/train-00002-of-00019.parquet
    - data/college_physics/train-00002-of-00018.parquet
    - data/college_biology/train-00002-of-00026.parquet
    - data/high_school_physics/train-00002-of-00011.parquet
    - data/conceptual_physics/train-00002-of-00012.parquet
    - data/high_school_statistics/train-00002-of-00017.parquet
    - data/astronomy/train-00002-of-00018.parquet
    - data/high_school_chemistry/train-00002-of-00018.parquet
    - data/econometrics/train-00002-of-00018.parquet
    - data/machine_learning/train-00002-of-00020.parquet
    - data/high_school_biology/train-00002-of-00032.parquet
    - data/college_computer_science/train-00002-of-00019.parquet
    - data/astronomy/train-00003-of-00018.parquet
    - data/high_school_chemistry/train-00003-of-00018.parquet
    - data/high_school_physics/train-00003-of-00011.parquet
    - data/geography/train-00003-of-00019.parquet
    - data/econometrics/train-00003-of-00018.parquet
    - data/professional_medicine/train-00003-of-00012.parquet
    - data/high_school_biology/train-00003-of-00032.parquet
    - data/college_medicine/train-00003-of-00040.parquet
    - data/high_school_statistics/train-00003-of-00017.parquet
    - data/college_maths/train-00003-of-00026.parquet
    - data/college_physics/train-00003-of-00018.parquet
    - data/high_school_maths/train-00003-of-00016.parquet
    - data/college_computer_science/train-00003-of-00019.parquet
    - data/high_school_computer_science/train-00003-of-00017.parquet
    - data/machine_learning/train-00003-of-00020.parquet
    - data/electronic_science/train-00003-of-00019.parquet
    - data/conceptual_physics/train-00003-of-00012.parquet
    - data/college_biology/train-00003-of-00026.parquet
    - data/college_chemistry/train-00003-of-00019.parquet
    - data/high_school_chemistry/train-00004-of-00018.parquet
    - data/high_school_biology/train-00004-of-00032.parquet
    - data/high_school_computer_science/train-00004-of-00017.parquet
    - data/electronic_science/train-00004-of-00019.parquet
    - data/professional_medicine/train-00004-of-00012.parquet
    - data/high_school_maths/train-00004-of-00016.parquet
    - data/college_biology/train-00004-of-00026.parquet
    - data/astronomy/train-00004-of-00018.parquet
    - data/high_school_physics/train-00004-of-00011.parquet
    - data/college_computer_science/train-00004-of-00019.parquet
    - data/econometrics/train-00004-of-00018.parquet
    - data/geography/train-00004-of-00019.parquet
    - data/college_maths/train-00004-of-00026.parquet
    - data/college_physics/train-00004-of-00018.parquet
    - data/high_school_statistics/train-00004-of-00017.parquet
    - data/machine_learning/train-00004-of-00020.parquet
    - data/conceptual_physics/train-00004-of-00012.parquet
    - data/college_medicine/train-00004-of-00040.parquet
    - data/college_chemistry/train-00004-of-00019.parquet
    - data/electronic_science/train-00005-of-00019.parquet
    - data/astronomy/train-00005-of-00018.parquet
    - data/college_biology/train-00005-of-00026.parquet
    - data/high_school_maths/train-00005-of-00016.parquet
    - data/econometrics/train-00005-of-00018.parquet
    - data/high_school_chemistry/train-00005-of-00018.parquet
    - data/college_medicine/train-00005-of-00040.parquet
    - data/high_school_computer_science/train-00005-of-00017.parquet
    - data/college_computer_science/train-00005-of-00019.parquet
    - data/high_school_biology/train-00005-of-00032.parquet
    - data/college_maths/train-00005-of-00026.parquet
    - data/college_chemistry/train-00005-of-00019.parquet
    - data/conceptual_physics/train-00005-of-00012.parquet
    - data/machine_learning/train-00005-of-00020.parquet
    - data/high_school_statistics/train-00005-of-00017.parquet
    - data/high_school_physics/train-00005-of-00011.parquet
    - data/professional_medicine/train-00005-of-00012.parquet
    - data/geography/train-00005-of-00019.parquet
    - data/college_physics/train-00005-of-00018.parquet
    - data/high_school_computer_science/train-00006-of-00017.parquet
    - data/conceptual_physics/train-00006-of-00012.parquet
    - data/electronic_science/train-00006-of-00019.parquet
    - data/college_medicine/train-00006-of-00040.parquet
    - data/professional_medicine/train-00006-of-00012.parquet
    - data/econometrics/train-00006-of-00018.parquet
    - data/college_computer_science/train-00006-of-00019.parquet
    - data/high_school_maths/train-00006-of-00016.parquet
    - data/college_physics/train-00006-of-00018.parquet
    - data/high_school_biology/train-00006-of-00032.parquet
    - data/geography/train-00006-of-00019.parquet
    - data/machine_learning/train-00006-of-00020.parquet
    - data/college_biology/train-00006-of-00026.parquet
    - data/astronomy/train-00006-of-00018.parquet
    - data/high_school_chemistry/train-00006-of-00018.parquet
    - data/college_chemistry/train-00006-of-00019.parquet
    - data/high_school_physics/train-00006-of-00011.parquet
    - data/high_school_statistics/train-00006-of-00017.parquet
    - data/college_maths/train-00006-of-00026.parquet
    - data/high_school_statistics/train-00007-of-00017.parquet
    - data/electronic_science/train-00007-of-00019.parquet
    - data/high_school_computer_science/train-00007-of-00017.parquet
    - data/high_school_biology/train-00007-of-00032.parquet
    - data/high_school_physics/train-00007-of-00011.parquet
    - data/high_school_maths/train-00007-of-00016.parquet
    - data/professional_medicine/train-00007-of-00012.parquet
    - data/conceptual_physics/train-00007-of-00012.parquet
    - data/college_chemistry/train-00007-of-00019.parquet
    - data/college_physics/train-00007-of-00018.parquet
    - data/college_maths/train-00007-of-00026.parquet
    - data/college_biology/train-00007-of-00026.parquet
    - data/college_medicine/train-00007-of-00040.parquet
    - data/geography/train-00007-of-00019.parquet
    - data/high_school_chemistry/train-00007-of-00018.parquet
    - data/econometrics/train-00007-of-00018.parquet
    - data/college_computer_science/train-00007-of-00019.parquet
    - data/astronomy/train-00007-of-00018.parquet
    - data/machine_learning/train-00007-of-00020.parquet
    - data/college_biology/train-00008-of-00026.parquet
    - data/college_computer_science/train-00008-of-00019.parquet
    - data/geography/train-00008-of-00019.parquet
    - data/high_school_physics/train-00008-of-00011.parquet
    - data/high_school_computer_science/train-00008-of-00017.parquet
    - data/machine_learning/train-00008-of-00020.parquet
    - data/professional_medicine/train-00008-of-00012.parquet
    - data/conceptual_physics/train-00008-of-00012.parquet
    - data/high_school_maths/train-00008-of-00016.parquet
    - data/astronomy/train-00008-of-00018.parquet
    - data/high_school_biology/train-00008-of-00032.parquet
    - data/econometrics/train-00008-of-00018.parquet
    - data/college_chemistry/train-00008-of-00019.parquet
    - data/high_school_chemistry/train-00008-of-00018.parquet
    - data/college_medicine/train-00008-of-00040.parquet
    - data/college_maths/train-00008-of-00026.parquet
    - data/college_physics/train-00008-of-00018.parquet
    - data/high_school_statistics/train-00008-of-00017.parquet
    - data/electronic_science/train-00008-of-00019.parquet
    - data/high_school_maths/train-00009-of-00016.parquet
    - data/college_physics/train-00009-of-00018.parquet
    - data/astronomy/train-00009-of-00018.parquet
    - data/college_computer_science/train-00009-of-00019.parquet
    - data/college_medicine/train-00009-of-00040.parquet
    - data/college_maths/train-00009-of-00026.parquet
    - data/high_school_biology/train-00009-of-00032.parquet
    - data/electronic_science/train-00009-of-00019.parquet
    - data/professional_medicine/train-00009-of-00012.parquet
    - data/geography/train-00009-of-00019.parquet
    - data/high_school_statistics/train-00009-of-00017.parquet
    - data/econometrics/train-00009-of-00018.parquet
    - data/high_school_chemistry/train-00009-of-00018.parquet
    - data/college_biology/train-00009-of-00026.parquet
    - data/high_school_computer_science/train-00009-of-00017.parquet
    - data/high_school_physics/train-00009-of-00011.parquet
    - data/machine_learning/train-00009-of-00020.parquet
    - data/college_chemistry/train-00009-of-00019.parquet
    - data/conceptual_physics/train-00009-of-00012.parquet
    - data/professional_medicine/train-00010-of-00012.parquet
    - data/college_physics/train-00010-of-00018.parquet
    - data/college_chemistry/train-00010-of-00019.parquet
    - data/high_school_computer_science/train-00010-of-00017.parquet
    - data/machine_learning/train-00010-of-00020.parquet
    - data/college_maths/train-00010-of-00026.parquet
    - data/astronomy/train-00010-of-00018.parquet
    - data/college_medicine/train-00010-of-00040.parquet
    - data/high_school_statistics/train-00010-of-00017.parquet
    - data/geography/train-00010-of-00019.parquet
    - data/college_computer_science/train-00010-of-00019.parquet
    - data/college_biology/train-00010-of-00026.parquet
    - data/conceptual_physics/train-00010-of-00012.parquet
    - data/econometrics/train-00010-of-00018.parquet
    - data/high_school_physics/train-00010-of-00011.parquet
    - data/high_school_maths/train-00010-of-00016.parquet
    - data/high_school_biology/train-00010-of-00032.parquet
    - data/high_school_chemistry/train-00010-of-00018.parquet
    - data/electronic_science/train-00010-of-00019.parquet
    - data/high_school_chemistry/train-00011-of-00018.parquet
    - data/college_physics/train-00011-of-00018.parquet
    - data/machine_learning/train-00011-of-00020.parquet
    - data/conceptual_physics/train-00011-of-00012.parquet
    - data/geography/train-00011-of-00019.parquet
    - data/electronic_science/train-00011-of-00019.parquet
    - data/high_school_biology/train-00011-of-00032.parquet
    - data/college_maths/train-00011-of-00026.parquet
    - data/high_school_maths/train-00011-of-00016.parquet
    - data/econometrics/train-00011-of-00018.parquet
    - data/high_school_computer_science/train-00011-of-00017.parquet
    - data/college_medicine/train-00011-of-00040.parquet
    - data/professional_medicine/train-00011-of-00012.parquet
    - data/high_school_statistics/train-00011-of-00017.parquet
    - data/college_computer_science/train-00011-of-00019.parquet
    - data/college_biology/train-00011-of-00026.parquet
    - data/college_chemistry/train-00011-of-00019.parquet
    - data/astronomy/train-00011-of-00018.parquet
    - data/electronic_science/train-00012-of-00019.parquet
    - data/astronomy/train-00012-of-00018.parquet
    - data/high_school_maths/train-00012-of-00016.parquet
    - data/college_maths/train-00012-of-00026.parquet
    - data/college_medicine/train-00012-of-00040.parquet
    - data/high_school_biology/train-00012-of-00032.parquet
    - data/college_chemistry/train-00012-of-00019.parquet
    - data/high_school_chemistry/train-00012-of-00018.parquet
    - data/geography/train-00012-of-00019.parquet
    - data/college_biology/train-00012-of-00026.parquet
    - data/college_computer_science/train-00012-of-00019.parquet
    - data/college_physics/train-00012-of-00018.parquet
    - data/high_school_statistics/train-00012-of-00017.parquet
    - data/machine_learning/train-00012-of-00020.parquet
    - data/high_school_computer_science/train-00012-of-00017.parquet
    - data/econometrics/train-00012-of-00018.parquet
    - data/college_chemistry/train-00013-of-00019.parquet
    - data/high_school_maths/train-00013-of-00016.parquet
    - data/high_school_statistics/train-00013-of-00017.parquet
    - data/high_school_computer_science/train-00013-of-00017.parquet
    - data/high_school_biology/train-00013-of-00032.parquet
    - data/econometrics/train-00013-of-00018.parquet
    - data/college_maths/train-00013-of-00026.parquet
    - data/electronic_science/train-00013-of-00019.parquet
    - data/geography/train-00013-of-00019.parquet
    - data/machine_learning/train-00013-of-00020.parquet
    - data/college_biology/train-00013-of-00026.parquet
    - data/college_medicine/train-00013-of-00040.parquet
    - data/high_school_chemistry/train-00013-of-00018.parquet
    - data/college_physics/train-00013-of-00018.parquet
    - data/astronomy/train-00013-of-00018.parquet
    - data/college_computer_science/train-00013-of-00019.parquet
    - data/college_physics/train-00014-of-00018.parquet
    - data/college_chemistry/train-00014-of-00019.parquet
    - data/college_maths/train-00014-of-00026.parquet
    - data/machine_learning/train-00014-of-00020.parquet
    - data/college_medicine/train-00014-of-00040.parquet
    - data/geography/train-00014-of-00019.parquet
    - data/high_school_chemistry/train-00014-of-00018.parquet
    - data/high_school_biology/train-00014-of-00032.parquet
    - data/astronomy/train-00014-of-00018.parquet
    - data/high_school_computer_science/train-00014-of-00017.parquet
    - data/electronic_science/train-00014-of-00019.parquet
    - data/econometrics/train-00014-of-00018.parquet
    - data/college_biology/train-00014-of-00026.parquet
    - data/high_school_maths/train-00014-of-00016.parquet
    - data/college_computer_science/train-00014-of-00019.parquet
    - data/high_school_statistics/train-00014-of-00017.parquet
    - data/geography/train-00015-of-00019.parquet
    - data/college_biology/train-00015-of-00026.parquet
    - data/econometrics/train-00015-of-00018.parquet
    - data/college_medicine/train-00015-of-00040.parquet
    - data/high_school_chemistry/train-00015-of-00018.parquet
    - data/college_computer_science/train-00015-of-00019.parquet
    - data/electronic_science/train-00015-of-00019.parquet
    - data/college_maths/train-00015-of-00026.parquet
    - data/high_school_biology/train-00015-of-00032.parquet
    - data/high_school_computer_science/train-00015-of-00017.parquet
    - data/high_school_statistics/train-00015-of-00017.parquet
    - data/college_chemistry/train-00015-of-00019.parquet
    - data/astronomy/train-00015-of-00018.parquet
    - data/machine_learning/train-00015-of-00020.parquet
    - data/high_school_maths/train-00015-of-00016.parquet
    - data/college_physics/train-00015-of-00018.parquet
    - data/high_school_chemistry/train-00016-of-00018.parquet
    - data/college_biology/train-00016-of-00026.parquet
    - data/college_physics/train-00016-of-00018.parquet
    - data/high_school_computer_science/train-00016-of-00017.parquet
    - data/econometrics/train-00016-of-00018.parquet
    - data/machine_learning/train-00016-of-00020.parquet
    - data/college_medicine/train-00016-of-00040.parquet
    - data/high_school_biology/train-00016-of-00032.parquet
    - data/electronic_science/train-00016-of-00019.parquet
    - data/geography/train-00016-of-00019.parquet
    - data/astronomy/train-00016-of-00018.parquet
    - data/high_school_statistics/train-00016-of-00017.parquet
    - data/college_maths/train-00016-of-00026.parquet
    - data/college_computer_science/train-00016-of-00019.parquet
    - data/college_chemistry/train-00016-of-00019.parquet
    - data/college_medicine/train-00017-of-00040.parquet
    - data/college_computer_science/train-00017-of-00019.parquet
    - data/high_school_biology/train-00017-of-00032.parquet
    - data/college_maths/train-00017-of-00026.parquet
    - data/geography/train-00017-of-00019.parquet
    - data/econometrics/train-00017-of-00018.parquet
    - data/college_physics/train-00017-of-00018.parquet
    - data/college_biology/train-00017-of-00026.parquet
    - data/high_school_chemistry/train-00017-of-00018.parquet
    - data/astronomy/train-00017-of-00018.parquet
    - data/college_chemistry/train-00017-of-00019.parquet
    - data/machine_learning/train-00017-of-00020.parquet
    - data/electronic_science/train-00017-of-00019.parquet
    - data/electronic_science/train-00018-of-00019.parquet
    - data/college_maths/train-00018-of-00026.parquet
    - data/college_computer_science/train-00018-of-00019.parquet
    - data/college_biology/train-00018-of-00026.parquet
    - data/machine_learning/train-00018-of-00020.parquet
    - data/high_school_biology/train-00018-of-00032.parquet
    - data/college_medicine/train-00018-of-00040.parquet
    - data/college_chemistry/train-00018-of-00019.parquet
    - data/geography/train-00018-of-00019.parquet
    - data/machine_learning/train-00019-of-00020.parquet
    - data/college_biology/train-00019-of-00026.parquet
    - data/college_maths/train-00019-of-00026.parquet
    - data/college_medicine/train-00019-of-00040.parquet
    - data/high_school_biology/train-00019-of-00032.parquet
    - data/high_school_biology/train-00020-of-00032.parquet
    - data/college_maths/train-00020-of-00026.parquet
    - data/college_biology/train-00020-of-00026.parquet
    - data/college_medicine/train-00020-of-00040.parquet
    - data/college_medicine/train-00021-of-00040.parquet
    - data/college_maths/train-00021-of-00026.parquet
    - data/college_biology/train-00021-of-00026.parquet
    - data/high_school_biology/train-00021-of-00032.parquet
    - data/college_medicine/train-00022-of-00040.parquet
    - data/college_biology/train-00022-of-00026.parquet
    - data/college_maths/train-00022-of-00026.parquet
    - data/high_school_biology/train-00022-of-00032.parquet
    - data/college_medicine/train-00023-of-00040.parquet
    - data/high_school_biology/train-00023-of-00032.parquet
    - data/college_maths/train-00023-of-00026.parquet
    - data/college_biology/train-00023-of-00026.parquet
    - data/college_maths/train-00024-of-00026.parquet
    - data/high_school_biology/train-00024-of-00032.parquet
    - data/college_medicine/train-00024-of-00040.parquet
    - data/college_biology/train-00024-of-00026.parquet
    - data/high_school_biology/train-00025-of-00032.parquet
    - data/college_maths/train-00025-of-00026.parquet
    - data/college_biology/train-00025-of-00026.parquet
    - data/college_medicine/train-00025-of-00040.parquet
    - data/high_school_biology/train-00026-of-00032.parquet
    - data/college_medicine/train-00026-of-00040.parquet
    - data/high_school_biology/train-00027-of-00032.parquet
    - data/college_medicine/train-00027-of-00040.parquet
    - data/college_medicine/train-00028-of-00040.parquet
    - data/high_school_biology/train-00028-of-00032.parquet
    - data/college_medicine/train-00029-of-00040.parquet
    - data/high_school_biology/train-00029-of-00032.parquet
    - data/high_school_biology/train-00030-of-00032.parquet
    - data/college_medicine/train-00030-of-00040.parquet
    - data/college_medicine/train-00031-of-00040.parquet
    - data/high_school_biology/train-00031-of-00032.parquet
    - data/college_medicine/train-00032-of-00040.parquet
    - data/college_medicine/train-00033-of-00040.parquet
    - data/college_medicine/train-00034-of-00040.parquet
    - data/college_medicine/train-00035-of-00040.parquet
    - data/college_medicine/train-00036-of-00040.parquet
    - data/college_medicine/train-00037-of-00040.parquet
    - data/college_medicine/train-00038-of-00040.parquet
    - data/college_medicine/train-00039-of-00040.parquet
- config_name: astronomy
  data_files:
  - split: train
    path:
    - data/astronomy/preview.parquet
    - data/astronomy/train-00000-of-00018.parquet
    - data/astronomy/train-00001-of-00018.parquet
    - data/astronomy/train-00002-of-00018.parquet
    - data/astronomy/train-00003-of-00018.parquet
    - data/astronomy/train-00004-of-00018.parquet
    - data/astronomy/train-00005-of-00018.parquet
    - data/astronomy/train-00006-of-00018.parquet
    - data/astronomy/train-00007-of-00018.parquet
    - data/astronomy/train-00008-of-00018.parquet
    - data/astronomy/train-00009-of-00018.parquet
    - data/astronomy/train-00010-of-00018.parquet
    - data/astronomy/train-00011-of-00018.parquet
    - data/astronomy/train-00012-of-00018.parquet
    - data/astronomy/train-00013-of-00018.parquet
    - data/astronomy/train-00014-of-00018.parquet
    - data/astronomy/train-00015-of-00018.parquet
    - data/astronomy/train-00016-of-00018.parquet
    - data/astronomy/train-00017-of-00018.parquet
- config_name: college_biology
  data_files:
  - split: train
    path:
    - data/college_biology/preview.parquet
    - data/college_biology/train-00000-of-00026.parquet
    - data/college_biology/train-00001-of-00026.parquet
    - data/college_biology/train-00002-of-00026.parquet
    - data/college_biology/train-00003-of-00026.parquet
    - data/college_biology/train-00004-of-00026.parquet
    - data/college_biology/train-00005-of-00026.parquet
    - data/college_biology/train-00006-of-00026.parquet
    - data/college_biology/train-00007-of-00026.parquet
    - data/college_biology/train-00008-of-00026.parquet
    - data/college_biology/train-00009-of-00026.parquet
    - data/college_biology/train-00010-of-00026.parquet
    - data/college_biology/train-00011-of-00026.parquet
    - data/college_biology/train-00012-of-00026.parquet
    - data/college_biology/train-00013-of-00026.parquet
    - data/college_biology/train-00014-of-00026.parquet
    - data/college_biology/train-00015-of-00026.parquet
    - data/college_biology/train-00016-of-00026.parquet
    - data/college_biology/train-00017-of-00026.parquet
    - data/college_biology/train-00018-of-00026.parquet
    - data/college_biology/train-00019-of-00026.parquet
    - data/college_biology/train-00020-of-00026.parquet
    - data/college_biology/train-00021-of-00026.parquet
    - data/college_biology/train-00022-of-00026.parquet
    - data/college_biology/train-00023-of-00026.parquet
    - data/college_biology/train-00024-of-00026.parquet
    - data/college_biology/train-00025-of-00026.parquet
- config_name: college_chemistry
  data_files:
  - split: train
    path:
    - data/college_chemistry/preview.parquet
    - data/college_chemistry/train-00000-of-00019.parquet
    - data/college_chemistry/train-00001-of-00019.parquet
    - data/college_chemistry/train-00002-of-00019.parquet
    - data/college_chemistry/train-00003-of-00019.parquet
    - data/college_chemistry/train-00004-of-00019.parquet
    - data/college_chemistry/train-00005-of-00019.parquet
    - data/college_chemistry/train-00006-of-00019.parquet
    - data/college_chemistry/train-00007-of-00019.parquet
    - data/college_chemistry/train-00008-of-00019.parquet
    - data/college_chemistry/train-00009-of-00019.parquet
    - data/college_chemistry/train-00010-of-00019.parquet
    - data/college_chemistry/train-00011-of-00019.parquet
    - data/college_chemistry/train-00012-of-00019.parquet
    - data/college_chemistry/train-00013-of-00019.parquet
    - data/college_chemistry/train-00014-of-00019.parquet
    - data/college_chemistry/train-00015-of-00019.parquet
    - data/college_chemistry/train-00016-of-00019.parquet
    - data/college_chemistry/train-00017-of-00019.parquet
    - data/college_chemistry/train-00018-of-00019.parquet
- config_name: college_computer_science
  data_files:
  - split: train
    path:
    - data/college_computer_science/preview.parquet
    - data/college_computer_science/train-00000-of-00019.parquet
    - data/college_computer_science/train-00001-of-00019.parquet
    - data/college_computer_science/train-00002-of-00019.parquet
    - data/college_computer_science/train-00003-of-00019.parquet
    - data/college_computer_science/train-00004-of-00019.parquet
    - data/college_computer_science/train-00005-of-00019.parquet
    - data/college_computer_science/train-00006-of-00019.parquet
    - data/college_computer_science/train-00007-of-00019.parquet
    - data/college_computer_science/train-00008-of-00019.parquet
    - data/college_computer_science/train-00009-of-00019.parquet
    - data/college_computer_science/train-00010-of-00019.parquet
    - data/college_computer_science/train-00011-of-00019.parquet
    - data/college_computer_science/train-00012-of-00019.parquet
    - data/college_computer_science/train-00013-of-00019.parquet
    - data/college_computer_science/train-00014-of-00019.parquet
    - data/college_computer_science/train-00015-of-00019.parquet
    - data/college_computer_science/train-00016-of-00019.parquet
    - data/college_computer_science/train-00017-of-00019.parquet
    - data/college_computer_science/train-00018-of-00019.parquet
- config_name: college_maths
  data_files:
  - split: train
    path:
    - data/college_maths/preview.parquet
    - data/college_maths/train-00000-of-00026.parquet
    - data/college_maths/train-00001-of-00026.parquet
    - data/college_maths/train-00002-of-00026.parquet
    - data/college_maths/train-00003-of-00026.parquet
    - data/college_maths/train-00004-of-00026.parquet
    - data/college_maths/train-00005-of-00026.parquet
    - data/college_maths/train-00006-of-00026.parquet
    - data/college_maths/train-00007-of-00026.parquet
    - data/college_maths/train-00008-of-00026.parquet
    - data/college_maths/train-00009-of-00026.parquet
    - data/college_maths/train-00010-of-00026.parquet
    - data/college_maths/train-00011-of-00026.parquet
    - data/college_maths/train-00012-of-00026.parquet
    - data/college_maths/train-00013-of-00026.parquet
    - data/college_maths/train-00014-of-00026.parquet
    - data/college_maths/train-00015-of-00026.parquet
    - data/college_maths/train-00016-of-00026.parquet
    - data/college_maths/train-00017-of-00026.parquet
    - data/college_maths/train-00018-of-00026.parquet
    - data/college_maths/train-00019-of-00026.parquet
    - data/college_maths/train-00020-of-00026.parquet
    - data/college_maths/train-00021-of-00026.parquet
    - data/college_maths/train-00022-of-00026.parquet
    - data/college_maths/train-00023-of-00026.parquet
    - data/college_maths/train-00024-of-00026.parquet
    - data/college_maths/train-00025-of-00026.parquet
- config_name: college_medicine
  data_files:
  - split: train
    path:
    - data/college_medicine/preview.parquet
    - data/college_medicine/train-00000-of-00040.parquet
    - data/college_medicine/train-00001-of-00040.parquet
    - data/college_medicine/train-00002-of-00040.parquet
    - data/college_medicine/train-00003-of-00040.parquet
    - data/college_medicine/train-00004-of-00040.parquet
    - data/college_medicine/train-00005-of-00040.parquet
    - data/college_medicine/train-00006-of-00040.parquet
    - data/college_medicine/train-00007-of-00040.parquet
    - data/college_medicine/train-00008-of-00040.parquet
    - data/college_medicine/train-00009-of-00040.parquet
    - data/college_medicine/train-00010-of-00040.parquet
    - data/college_medicine/train-00011-of-00040.parquet
    - data/college_medicine/train-00012-of-00040.parquet
    - data/college_medicine/train-00013-of-00040.parquet
    - data/college_medicine/train-00014-of-00040.parquet
    - data/college_medicine/train-00015-of-00040.parquet
    - data/college_medicine/train-00016-of-00040.parquet
    - data/college_medicine/train-00017-of-00040.parquet
    - data/college_medicine/train-00018-of-00040.parquet
    - data/college_medicine/train-00019-of-00040.parquet
    - data/college_medicine/train-00020-of-00040.parquet
    - data/college_medicine/train-00021-of-00040.parquet
    - data/college_medicine/train-00022-of-00040.parquet
    - data/college_medicine/train-00023-of-00040.parquet
    - data/college_medicine/train-00024-of-00040.parquet
    - data/college_medicine/train-00025-of-00040.parquet
    - data/college_medicine/train-00026-of-00040.parquet
    - data/college_medicine/train-00027-of-00040.parquet
    - data/college_medicine/train-00028-of-00040.parquet
    - data/college_medicine/train-00029-of-00040.parquet
    - data/college_medicine/train-00030-of-00040.parquet
    - data/college_medicine/train-00031-of-00040.parquet
    - data/college_medicine/train-00032-of-00040.parquet
    - data/college_medicine/train-00033-of-00040.parquet
    - data/college_medicine/train-00034-of-00040.parquet
    - data/college_medicine/train-00035-of-00040.parquet
    - data/college_medicine/train-00036-of-00040.parquet
    - data/college_medicine/train-00037-of-00040.parquet
    - data/college_medicine/train-00038-of-00040.parquet
    - data/college_medicine/train-00039-of-00040.parquet
- config_name: college_physics
  data_files:
  - split: train
    path:
    - data/college_physics/preview.parquet
    - data/college_physics/train-00000-of-00018.parquet
    - data/college_physics/train-00001-of-00018.parquet
    - data/college_physics/train-00002-of-00018.parquet
    - data/college_physics/train-00003-of-00018.parquet
    - data/college_physics/train-00004-of-00018.parquet
    - data/college_physics/train-00005-of-00018.parquet
    - data/college_physics/train-00006-of-00018.parquet
    - data/college_physics/train-00007-of-00018.parquet
    - data/college_physics/train-00008-of-00018.parquet
    - data/college_physics/train-00009-of-00018.parquet
    - data/college_physics/train-00010-of-00018.parquet
    - data/college_physics/train-00011-of-00018.parquet
    - data/college_physics/train-00012-of-00018.parquet
    - data/college_physics/train-00013-of-00018.parquet
    - data/college_physics/train-00014-of-00018.parquet
    - data/college_physics/train-00015-of-00018.parquet
    - data/college_physics/train-00016-of-00018.parquet
    - data/college_physics/train-00017-of-00018.parquet
- config_name: conceptual_physics
  data_files:
  - split: train
    path:
    - data/conceptual_physics/preview.parquet
    - data/conceptual_physics/train-00000-of-00012.parquet
    - data/conceptual_physics/train-00001-of-00012.parquet
    - data/conceptual_physics/train-00002-of-00012.parquet
    - data/conceptual_physics/train-00003-of-00012.parquet
    - data/conceptual_physics/train-00004-of-00012.parquet
    - data/conceptual_physics/train-00005-of-00012.parquet
    - data/conceptual_physics/train-00006-of-00012.parquet
    - data/conceptual_physics/train-00007-of-00012.parquet
    - data/conceptual_physics/train-00008-of-00012.parquet
    - data/conceptual_physics/train-00009-of-00012.parquet
    - data/conceptual_physics/train-00010-of-00012.parquet
    - data/conceptual_physics/train-00011-of-00012.parquet
- config_name: econometrics
  data_files:
  - split: train
    path:
    - data/econometrics/preview.parquet
    - data/econometrics/train-00000-of-00018.parquet
    - data/econometrics/train-00001-of-00018.parquet
    - data/econometrics/train-00002-of-00018.parquet
    - data/econometrics/train-00003-of-00018.parquet
    - data/econometrics/train-00004-of-00018.parquet
    - data/econometrics/train-00005-of-00018.parquet
    - data/econometrics/train-00006-of-00018.parquet
    - data/econometrics/train-00007-of-00018.parquet
    - data/econometrics/train-00008-of-00018.parquet
    - data/econometrics/train-00009-of-00018.parquet
    - data/econometrics/train-00010-of-00018.parquet
    - data/econometrics/train-00011-of-00018.parquet
    - data/econometrics/train-00012-of-00018.parquet
    - data/econometrics/train-00013-of-00018.parquet
    - data/econometrics/train-00014-of-00018.parquet
    - data/econometrics/train-00015-of-00018.parquet
    - data/econometrics/train-00016-of-00018.parquet
    - data/econometrics/train-00017-of-00018.parquet
- config_name: electronic_science
  data_files:
  - split: train
    path:
    - data/electronic_science/preview.parquet
    - data/electronic_science/train-00000-of-00019.parquet
    - data/electronic_science/train-00001-of-00019.parquet
    - data/electronic_science/train-00002-of-00019.parquet
    - data/electronic_science/train-00003-of-00019.parquet
    - data/electronic_science/train-00004-of-00019.parquet
    - data/electronic_science/train-00005-of-00019.parquet
    - data/electronic_science/train-00006-of-00019.parquet
    - data/electronic_science/train-00007-of-00019.parquet
    - data/electronic_science/train-00008-of-00019.parquet
    - data/electronic_science/train-00009-of-00019.parquet
    - data/electronic_science/train-00010-of-00019.parquet
    - data/electronic_science/train-00011-of-00019.parquet
    - data/electronic_science/train-00012-of-00019.parquet
    - data/electronic_science/train-00013-of-00019.parquet
    - data/electronic_science/train-00014-of-00019.parquet
    - data/electronic_science/train-00015-of-00019.parquet
    - data/electronic_science/train-00016-of-00019.parquet
    - data/electronic_science/train-00017-of-00019.parquet
    - data/electronic_science/train-00018-of-00019.parquet
- config_name: geography
  data_files:
  - split: train
    path:
    - data/geography/preview.parquet
    - data/geography/train-00000-of-00019.parquet
    - data/geography/train-00001-of-00019.parquet
    - data/geography/train-00002-of-00019.parquet
    - data/geography/train-00003-of-00019.parquet
    - data/geography/train-00004-of-00019.parquet
    - data/geography/train-00005-of-00019.parquet
    - data/geography/train-00006-of-00019.parquet
    - data/geography/train-00007-of-00019.parquet
    - data/geography/train-00008-of-00019.parquet
    - data/geography/train-00009-of-00019.parquet
    - data/geography/train-00010-of-00019.parquet
    - data/geography/train-00011-of-00019.parquet
    - data/geography/train-00012-of-00019.parquet
    - data/geography/train-00013-of-00019.parquet
    - data/geography/train-00014-of-00019.parquet
    - data/geography/train-00015-of-00019.parquet
    - data/geography/train-00016-of-00019.parquet
    - data/geography/train-00017-of-00019.parquet
    - data/geography/train-00018-of-00019.parquet
- config_name: high_school_biology
  data_files:
  - split: train
    path:
    - data/high_school_biology/preview.parquet
    - data/high_school_biology/train-00000-of-00032.parquet
    - data/high_school_biology/train-00001-of-00032.parquet
    - data/high_school_biology/train-00002-of-00032.parquet
    - data/high_school_biology/train-00003-of-00032.parquet
    - data/high_school_biology/train-00004-of-00032.parquet
    - data/high_school_biology/train-00005-of-00032.parquet
    - data/high_school_biology/train-00006-of-00032.parquet
    - data/high_school_biology/train-00007-of-00032.parquet
    - data/high_school_biology/train-00008-of-00032.parquet
    - data/high_school_biology/train-00009-of-00032.parquet
    - data/high_school_biology/train-00010-of-00032.parquet
    - data/high_school_biology/train-00011-of-00032.parquet
    - data/high_school_biology/train-00012-of-00032.parquet
    - data/high_school_biology/train-00013-of-00032.parquet
    - data/high_school_biology/train-00014-of-00032.parquet
    - data/high_school_biology/train-00015-of-00032.parquet
    - data/high_school_biology/train-00016-of-00032.parquet
    - data/high_school_biology/train-00017-of-00032.parquet
    - data/high_school_biology/train-00018-of-00032.parquet
    - data/high_school_biology/train-00019-of-00032.parquet
    - data/high_school_biology/train-00020-of-00032.parquet
    - data/high_school_biology/train-00021-of-00032.parquet
    - data/high_school_biology/train-00022-of-00032.parquet
    - data/high_school_biology/train-00023-of-00032.parquet
    - data/high_school_biology/train-00024-of-00032.parquet
    - data/high_school_biology/train-00025-of-00032.parquet
    - data/high_school_biology/train-00026-of-00032.parquet
    - data/high_school_biology/train-00027-of-00032.parquet
    - data/high_school_biology/train-00028-of-00032.parquet
    - data/high_school_biology/train-00029-of-00032.parquet
    - data/high_school_biology/train-00030-of-00032.parquet
    - data/high_school_biology/train-00031-of-00032.parquet
- config_name: high_school_chemistry
  data_files:
  - split: train
    path:
    - data/high_school_chemistry/preview.parquet
    - data/high_school_chemistry/train-00000-of-00018.parquet
    - data/high_school_chemistry/train-00001-of-00018.parquet
    - data/high_school_chemistry/train-00002-of-00018.parquet
    - data/high_school_chemistry/train-00003-of-00018.parquet
    - data/high_school_chemistry/train-00004-of-00018.parquet
    - data/high_school_chemistry/train-00005-of-00018.parquet
    - data/high_school_chemistry/train-00006-of-00018.parquet
    - data/high_school_chemistry/train-00007-of-00018.parquet
    - data/high_school_chemistry/train-00008-of-00018.parquet
    - data/high_school_chemistry/train-00009-of-00018.parquet
    - data/high_school_chemistry/train-00010-of-00018.parquet
    - data/high_school_chemistry/train-00011-of-00018.parquet
    - data/high_school_chemistry/train-00012-of-00018.parquet
    - data/high_school_chemistry/train-00013-of-00018.parquet
    - data/high_school_chemistry/train-00014-of-00018.parquet
    - data/high_school_chemistry/train-00015-of-00018.parquet
    - data/high_school_chemistry/train-00016-of-00018.parquet
    - data/high_school_chemistry/train-00017-of-00018.parquet
- config_name: high_school_computer_science
  data_files:
  - split: train
    path:
    - data/high_school_computer_science/preview.parquet
    - data/high_school_computer_science/train-00000-of-00017.parquet
    - data/high_school_computer_science/train-00001-of-00017.parquet
    - data/high_school_computer_science/train-00002-of-00017.parquet
    - data/high_school_computer_science/train-00003-of-00017.parquet
    - data/high_school_computer_science/train-00004-of-00017.parquet
    - data/high_school_computer_science/train-00005-of-00017.parquet
    - data/high_school_computer_science/train-00006-of-00017.parquet
    - data/high_school_computer_science/train-00007-of-00017.parquet
    - data/high_school_computer_science/train-00008-of-00017.parquet
    - data/high_school_computer_science/train-00009-of-00017.parquet
    - data/high_school_computer_science/train-00010-of-00017.parquet
    - data/high_school_computer_science/train-00011-of-00017.parquet
    - data/high_school_computer_science/train-00012-of-00017.parquet
    - data/high_school_computer_science/train-00013-of-00017.parquet
    - data/high_school_computer_science/train-00014-of-00017.parquet
    - data/high_school_computer_science/train-00015-of-00017.parquet
    - data/high_school_computer_science/train-00016-of-00017.parquet
- config_name: high_school_maths
  data_files:
  - split: train
    path:
    - data/high_school_maths/preview.parquet
    - data/high_school_maths/train-00000-of-00016.parquet
    - data/high_school_maths/train-00001-of-00016.parquet
    - data/high_school_maths/train-00002-of-00016.parquet
    - data/high_school_maths/train-00003-of-00016.parquet
    - data/high_school_maths/train-00004-of-00016.parquet
    - data/high_school_maths/train-00005-of-00016.parquet
    - data/high_school_maths/train-00006-of-00016.parquet
    - data/high_school_maths/train-00007-of-00016.parquet
    - data/high_school_maths/train-00008-of-00016.parquet
    - data/high_school_maths/train-00009-of-00016.parquet
    - data/high_school_maths/train-00010-of-00016.parquet
    - data/high_school_maths/train-00011-of-00016.parquet
    - data/high_school_maths/train-00012-of-00016.parquet
    - data/high_school_maths/train-00013-of-00016.parquet
    - data/high_school_maths/train-00014-of-00016.parquet
    - data/high_school_maths/train-00015-of-00016.parquet
- config_name: high_school_physics
  data_files:
  - split: train
    path:
    - data/high_school_physics/preview.parquet
    - data/high_school_physics/train-00000-of-00011.parquet
    - data/high_school_physics/train-00001-of-00011.parquet
    - data/high_school_physics/train-00002-of-00011.parquet
    - data/high_school_physics/train-00003-of-00011.parquet
    - data/high_school_physics/train-00004-of-00011.parquet
    - data/high_school_physics/train-00005-of-00011.parquet
    - data/high_school_physics/train-00006-of-00011.parquet
    - data/high_school_physics/train-00007-of-00011.parquet
    - data/high_school_physics/train-00008-of-00011.parquet
    - data/high_school_physics/train-00009-of-00011.parquet
    - data/high_school_physics/train-00010-of-00011.parquet
- config_name: high_school_statistics
  data_files:
  - split: train
    path:
    - data/high_school_statistics/preview.parquet
    - data/high_school_statistics/train-00000-of-00017.parquet
    - data/high_school_statistics/train-00001-of-00017.parquet
    - data/high_school_statistics/train-00002-of-00017.parquet
    - data/high_school_statistics/train-00003-of-00017.parquet
    - data/high_school_statistics/train-00004-of-00017.parquet
    - data/high_school_statistics/train-00005-of-00017.parquet
    - data/high_school_statistics/train-00006-of-00017.parquet
    - data/high_school_statistics/train-00007-of-00017.parquet
    - data/high_school_statistics/train-00008-of-00017.parquet
    - data/high_school_statistics/train-00009-of-00017.parquet
    - data/high_school_statistics/train-00010-of-00017.parquet
    - data/high_school_statistics/train-00011-of-00017.parquet
    - data/high_school_statistics/train-00012-of-00017.parquet
    - data/high_school_statistics/train-00013-of-00017.parquet
    - data/high_school_statistics/train-00014-of-00017.parquet
    - data/high_school_statistics/train-00015-of-00017.parquet
    - data/high_school_statistics/train-00016-of-00017.parquet
- config_name: machine_learning
  data_files:
  - split: train
    path:
    - data/machine_learning/preview.parquet
    - data/machine_learning/train-00000-of-00020.parquet
    - data/machine_learning/train-00001-of-00020.parquet
    - data/machine_learning/train-00002-of-00020.parquet
    - data/machine_learning/train-00003-of-00020.parquet
    - data/machine_learning/train-00004-of-00020.parquet
    - data/machine_learning/train-00005-of-00020.parquet
    - data/machine_learning/train-00006-of-00020.parquet
    - data/machine_learning/train-00007-of-00020.parquet
    - data/machine_learning/train-00008-of-00020.parquet
    - data/machine_learning/train-00009-of-00020.parquet
    - data/machine_learning/train-00010-of-00020.parquet
    - data/machine_learning/train-00011-of-00020.parquet
    - data/machine_learning/train-00012-of-00020.parquet
    - data/machine_learning/train-00013-of-00020.parquet
    - data/machine_learning/train-00014-of-00020.parquet
    - data/machine_learning/train-00015-of-00020.parquet
    - data/machine_learning/train-00016-of-00020.parquet
    - data/machine_learning/train-00017-of-00020.parquet
    - data/machine_learning/train-00018-of-00020.parquet
    - data/machine_learning/train-00019-of-00020.parquet
- config_name: professional_medicine
  data_files:
  - split: train
    path:
    - data/professional_medicine/preview.parquet
    - data/professional_medicine/train-00000-of-00012.parquet
    - data/professional_medicine/train-00001-of-00012.parquet
    - data/professional_medicine/train-00002-of-00012.parquet
    - data/professional_medicine/train-00003-of-00012.parquet
    - data/professional_medicine/train-00004-of-00012.parquet
    - data/professional_medicine/train-00005-of-00012.parquet
    - data/professional_medicine/train-00006-of-00012.parquet
    - data/professional_medicine/train-00007-of-00012.parquet
    - data/professional_medicine/train-00008-of-00012.parquet
    - data/professional_medicine/train-00009-of-00012.parquet
    - data/professional_medicine/train-00010-of-00012.parquet
    - data/professional_medicine/train-00011-of-00012.parquet
---

# QVAC Genesis III

This repository contains the QVAC Genesis III dataset associated with the
COLM 2026 paper
[QVAC Genesis III: A Large-Scale, High-Quality Open Synthetic STEM Corpus for Efficient Language Model Pre-Training](https://arxiv.org/abs/2609.19513).

| Resource | Link |
| -------- | ---- |
| 📄 Paper (COLM 2026) | [arXiv:2609.19513](https://arxiv.org/abs/2609.19513) |
| 📝 Blog | [huggingface.co/blog/qvac/genesis-iii](https://huggingface.co/blog/qvac/genesis-iii) |
| 💻 Code | [github.com/tether-ai-research/qvac-genesis-III](https://github.com/tether-ai-research/qvac-genesis-III) |
| 🤗 Models | [huggingface.co/collections/qvac/genesis-iii](https://huggingface.co/collections/qvac/genesis-iii) |

QVAC Genesis III is a large-scale synthetic STEM corpus for efficient language-model pre-training. It contains **159,646,553 educational documents** and approximately **191 billion training tokens** across 19 curriculum-aligned domains, two complementary generation pipelines, and four educational writing styles.

![QVAC Genesis III synthetic-data generation pipeline, from seed acquisition and multiple-choice question generation through student answering, LLM parsing, Option-Level Reasoning, and Failure Analysis](https://cdn-uploads.huggingface.co/production/uploads/66ad47f5a45133da70d1c40b/Ha8so8hEmzOpkPWHfbMo3.png)

## 📋 Dataset Details

| Field | Value |
| ----- | ----- |
| **Curated by** | Davide Vitabile, Nikhil Ranjan, Akshay Nambiar, Kamal Kumar Gupta, and Amril Nazir |
| **Organization** | Tether Data, S.A. de C.V. (Tether AI Research) |
| **Shared by** | QVAC |
| **Language** | English |
| **License** | CC BY-NC 4.0 |
| **Scale** | 159,646,553 documents and approximately 191 billion training tokens |
| **Domains** | 19 STEM domains |
| **Generation methods** | Failure Analysis and Option-Level Reasoning |
| **Styles** | educational textbook, web article, question-answer tutoring, and conversational dialogue |

## ✨ What Genesis III Adds

QVAC Genesis III uses both successful and unsuccessful answers from an edge-scale student model as generation signals:

- **Failure Analysis (`failure_analysis`)** converts incorrect or non-extractable student responses into corrective lessons. The generated document diagnoses the likely mistake, explains the correct approach, and restates the problem so the result is self-contained.
- **Option-Level Reasoning (`option_level`)** converts correctly answered questions into contrastive explanations that justify the correct option and explain why each distractor is wrong.

Both pipelines render their output in four styles to increase structural and pedagogical diversity.

Under controlled training settings, models trained from scratch on Failure Analysis, Option-Level Reasoning, and their combination substantially improve ARC, GPQA Diamond, and MMLU STEM performance over token-matched Cosmopedia-v2 baselines.

![QVAC Genesis III Combined compared with Cosmo-1B and the seven-epoch Cosmopedia-v2 baseline on ARC-Easy, ARC-Challenge, GPQA Diamond, MMLU STEM accuracy, and MMLU STEM Valid Answer Rate](https://cdn-uploads.huggingface.co/production/uploads/66ad47f5a45133da70d1c40b/eaGZc_nKwJcZai0JVmtOL.png)

## 🧪 Research Checkpoints

Models trained in the paper's controlled from-scratch experiments:

| Checkpoint | Role | Hugging Face |
| ---------- | ---- | ------------ |
| 🔍 Failure Analysis | FA data ablation | [qvac-genesis-iii-qwen3-1.7b-fa](https://huggingface.co/qvac/qvac-genesis-iii-qwen3-1.7b-fa) |
| 🧠 Option-Level | OL data ablation | [qvac-genesis-iii-qwen3-1.7b-ol](https://huggingface.co/qvac/qvac-genesis-iii-qwen3-1.7b-ol) |
| 🔗 Combined FA+OL | Full Genesis III mix | [qvac-genesis-iii-qwen3-1.7b-combined](https://huggingface.co/qvac/qvac-genesis-iii-qwen3-1.7b-combined) |
## 🗂️ Dataset Configurations

The dataset has one `train` split and 20 configurations:

- `full` (default): all 19 domains
- one configuration for each individual domain

The configurations reference the same underlying Parquet files; selecting a domain does not create a duplicate copy of the data.

### Domains

| | | |
| --- | --- | --- |
| `astronomy` | `college_biology` | `college_chemistry` |
| `college_computer_science` | `college_maths` | `college_medicine` |
| `college_physics` | `conceptual_physics` | `econometrics` |
| `electronic_science` | `geography` | `high_school_biology` |
| `high_school_chemistry` | `high_school_computer_science` | `high_school_maths` |
| `high_school_physics` | `high_school_statistics` | `machine_learning` |
| `professional_medicine` | | |

Release identifiers preserve two source-directory names:
`electronic_science` corresponds to **electrical engineering**, and
`geography` corresponds to **high school geography**.

## 🧱 Dataset Structure

Each row contains the final self-contained training document and explicit metadata:

| Field | Type | Description |
|---|---|---|
| `text` | string | Final synthetic educational document used for pre-training. |
| `domain` | string | One of the 19 curriculum-aligned STEM domains. |
| `pipeline` | string | `failure_analysis` or `option_level`. |
| `style` | string | `educational_textbook`, `web_article`, `qa`, or `conversational_dialogue`. |

Generation prompts, generator scratch reasoning, duplicated output fields, and the redundant correctness flag are not included in the public schema.
We applied the same record filter used for training, keeping documents whose
`text` is present, non-empty, and longer than 100 characters and whose
`reasoning_output` is present and valid. This filtering removed approximately
1.02 million incomplete or placeholder source records.

Illustrative, abridged example:

```json
{
  "text": "**User:** I have a question about geostationary orbit...",
  "domain": "astronomy",
  "pipeline": "option_level",
  "style": "conversational_dialogue"
}
```

## 🚀 How to Load

The full corpus is very large. Streaming is strongly recommended unless you intentionally want a local copy.

```python
from datasets import load_dataset

# Stream the complete corpus (the default configuration is "full").
dataset = load_dataset(
    "qvac/GenesisIII",
    split="train",
    streaming=True,
)
example = next(iter(dataset))
print(example)
```

Load one domain:

```python
from datasets import load_dataset

astronomy = load_dataset(
    "qvac/GenesisIII",
    "astronomy",
    split="train",
    streaming=True,
)
```

Filter by pipeline or style:

```python
failure_analysis = astronomy.filter(
    lambda row: row["pipeline"] == "failure_analysis"
)

textbook = astronomy.filter(
    lambda row: row["style"] == "educational_textbook"
)
```

## 🏗️ Dataset Creation

### 1. Seed acquisition and quality filtering

Seed passages were sampled from [FineFineWeb](https://huggingface.co/datasets/m-a-p/FineFineWeb) STEM categories and mapped to curriculum-aligned target domains. The [Ultra-FineWeb classifier](https://huggingface.co/openbmb/Ultra-FineWeb-classifier) retained high-quality passages.

### 2. Multiple-choice question generation

[QwQ-32B](https://huggingface.co/Qwen/QwQ-32B) generated self-contained multiple-choice questions with four mutually exclusive options and one gold answer. A format validator rejected malformed generations.

### 3. Student answering and answer extraction

[Qwen3-1.7B-Base](https://huggingface.co/Qwen/Qwen3-1.7B-Base) produced a free-form answer to each question. [CompassJudger-2-32B-Instruct](https://huggingface.co/opencompass/CompassJudger-2-32B-Instruct) acted as a parser: it extracted one final option or returned a non-answer/ambiguity outcome. It was used to recover the student's committed choice, not to solve the question.

### 4. Dual-route educational generation

Incorrect and non-extractable responses were routed to Failure Analysis. Correct responses were routed to Option-Level Reasoning. QwQ-32B generated self-contained documents in four styles:

- educational textbook;
- web article;
- question-answer tutoring;
- conversational dialogue.

The complete methodology and prompt templates are documented in the
[Genesis III publication](https://arxiv.org/abs/2609.19513).

## 🛡️ Content Safety and Personal Information

The complete corpus was scanned before
release. The scan covered all source text fields rather than a sample:

- The toxicity pipeline used **Detoxify** for broad screening and
  **Qwen3Guard-Gen-8B** for contextual review of flagged passages.
- The PII pipeline detected common personal-data categories, including
  email addresses, phone numbers, network addresses, physical addresses, and
  financial identifiers. **CompassJudger-2-32B-Instruct** reviewed ambiguous
  findings in context.
- Confirmed personal identifiers were removed by replacing them with safe,
  non-resolving synthetic placeholders. The cleaned corpus was then scanned
  again to verify that no confirmed identifiers remained and that its structure
  was intact.

The personal-information clean-up modified 379,336 of 160,668,773 source
records, approximately **0.24%**, removing confirmed identifiers through
placeholder substitution. Contextual review found **zero toxic passages**, so
no toxicity-based removals were necessary.

These automated and model-assisted scans can produce false positives or false
negatives. Redaction may also affect a small amount of non-personal text that
resembles an identifier. The scan addresses toxicity and personal information;
it does not establish factual correctness, copyright clearance, or suitability
for a particular downstream use. Although automated cleaning, filtering, and deduplication were applied, the dataset was not exhaustively audited for personal, sensitive, offensive, or otherwise harmful content. Users should perform additional screening appropriate to their use of the dataset.

## 🧹 Deduplication and Benchmark Decontamination

- MinHash near-deduplication with 60-gram signatures flagged 47,927 candidate pairs and identified 1,729 unique near-duplicate documents, less than 0.002% of the corpus.
- AllenAI Decon overlap checks identified 17 verified benchmark-contaminated documents across the corpus.
- All verified matches occurred in the Failure Analysis data; Option-Level Reasoning had no verified matches.
- GSM8K test leakage was 0%; MMLU test overlap was 0.005%.

These automatic and manually reviewed checks quantify known overlap but cannot prove that the corpus is free of all benchmark material. The retention or removal status of the identified near-duplicates and benchmark matches must be confirmed before publication.

## 🎯 Uses

### ✅ Direct Use

The dataset is intended for research and education purposes and in connection with those uses, for:


- pre-training and continual pre-training of language models;
- research on synthetic educational and STEM data;
- controlled studies of failure-focused and option-level supervision;
- curriculum construction by domain, pipeline, or educational style;
- research on token-efficient training of small language models.

### 🚫 Out-of-Scope Use

The dataset should not be treated as:

- a factual reference or source of ground truth;
- a replacement for textbooks, instructors, clinicians, engineers, or other domain experts;
- human-authored educational material;
- validated training data for medical, legal, safety-critical, or other high-stakes deployment;
- evidence that a downstream model is safe or reliable.

Models trained on this corpus require independent domain-specific and safety evaluation before deployment.

## ⚠️ Bias, Risks, and Limitations

- The documents are synthetic and may contain hallucinations, reasoning errors, contradictory explanations, or incorrect answer keys.
- Generator, student, parser, and filtering models can transmit their own biases and systematic errors.
- Coverage is restricted to English-language STEM and is not representative of broader languages, cultures, disciplines, or educational settings.
- Domain and difficulty distributions are uneven.
- Educational styles are generated templates, not naturally occurring classroom interactions.
- Automated PII and toxicity screening can produce both false positives and false negatives.
- Redaction placeholders can reduce fidelity or collapse distinct values into the same token.
- Benchmark decontamination is threshold-based and may miss semantic or paraphrased overlap.
- Strong benchmark results do not imply suitability for high-stakes applications.

Users should inspect and filter the corpus for their specific use case.

## ⚖️ Licensing

Licensing Information: This QVAC Genesis III dataset is licensed by Tether Data, S.A. de C.V. under the CC-BY-NC 4.0 (https://creativecommons.org/licenses/by-nc/4.0/legalcode.en). As described in the blog post (https://huggingface.co/blog/qvac/genesis-iii/), the data was generated using FineFineWeb (https://huggingface.co/datasets/m-a-p/FineFineWeb) licensed under the Apache 2.0 license, the Ultra-FineWeb-classifier (https://huggingface.co/openbmb/Ultra-FineWeb-classifier) licensed under the Apache 2.0 license, QwQ-32B (https://huggingface.co/Qwen/QwQ-32B) licensed under the Apache 2.0 license and Qwen/Qwen3-1.7B-Base (https://huggingface.co/Qwen/Qwen3-1.7B-Base) licensed under the Apache 2.0 license.

FineFineWeb data was generated from FineWeb data (ODC-BY 1.0) using data that originated in Common Crawl (Creative Commons terms). The FineFineWeb authors used Qwen2-7B licensed under the Apache-2.0 license, Qwen 2-72B-Instruct licensed under the Tongyi Qianwen license, and a BERT model licensed under the Apache 2.0 license.

Users are responsible for reviewing the applicable licenses and terms of these upstream resources. The dataset is provided as-is, without warranties of factual accuracy, fitness for a particular purpose, or non-infringement.

## 📮 Copyright Complaints

We will take appropriate action in response to notices of copyright infringement. If you believe your work has been used or copied in a way that infringes your intellectual-property rights, email [data-apps@tether.io](mailto:data-apps@tether.io) and identify the copyrighted work and the allegedly infringing material.

## 📖 Citation

```bibtex
@misc{vitabile2026qvacgenesisiii,
  title         = {QVAC Genesis III: A Large-Scale, High-Quality Open Synthetic STEM Corpus for Efficient Language Model Pre-Training},
  author        = {Davide Vitabile and Nikhil Ranjan and Akshay Nambiar and Kamal Kumar Gupta and Amril Nazir},
  year          = {2026},
  eprint        = {2609.19513},
  archivePrefix = {arXiv},
  primaryClass  = {cs.AI},
  institution   = {Tether Data, S.A. de C.V. d.b.a. Tether AI Research},
  note          = {Accepted at the Conference on Language Modeling (COLM) 2026},
  url           = {https://arxiv.org/abs/2609.19513}
}
```

## ✉️ Dataset Card Contact

Questions and feedback can be submitted through the [QVAC organization on Hugging Face](https://huggingface.co/qvac). Copyright complaints should be sent to [data-apps@tether.io](mailto:data-apps@tether.io).