LLM-Generated Java Programming MCQ Dataset with Bloom-Level Annotations

# LLM-Generated MCQ Dataset with Bloom-Level Annotations This dataset contains 5,070 multiple-choice questions generated by GPT-3.5 and Gemini across 13 Java programming concepts and 5 Bloom's Taxonomy levels. Each entry includes: - The question stem- Answer options (A–D)- Correct answer- Explanations for each option- Intended Bloom level (from prompt)- Inferred Bloom level (from Claude-3 evaluation)- Bloom alignment score- Explanation clarity, relevance, and answer accuracy scores- A mismatch flag for cognitive misalignment ## Format Provided as a single CSV file with the following fields:- `question`, `option_A`, ..., `option_D`, `correct_answer`- `bloom_level`, `inferred_bloom`, `bloom_mismatch`- `explanation`, `explainability_score`, `accuracy_score`, etc. ## License This dataset is shared under the [CC-BY 4.0](https://creativecommons.org/licenses/by/4.0/) license. ## Citation If you use this dataset, please cite: > Yetunde Folajimi. (2025). *Explainability-Driven Detection of Cognitive Misalignments in AI-Generated Educational Content* [Dataset]. Zenodo. https://doi.org/15238053 ## Contact For questions or collaboration, contact: [e-mail]