stair-lab/skill_composition_hypothesis

A Dataset for Skill Composition Hypothesis We employ a language model to annotate the specific skills assessed by each question derived from various Natural Language Processing benchmarks. The skill taxonomy utilized is sourced from IXL. The associated GitHub repository that produces this dataset can be found here: https://github.com/sangttruong/skill-composition-hypothesis.