Scaling Interdisciplinary Scoping Reviews with Large Language Models: A Validated Human-in-the-Loop Framework - Dataset and all calculations

This dataset accompanies a research paper introducing and validating an AI-assisted screening methodology for interdisciplinary scoping reviews, applied to the domain of AI, energy, and ethics in LMICs. It includes the literature search results, Python code for prompting six GPT models across three prompt levels, resulting model scores, and a final comparison with human reviewer benchmarks using a custom Weighted Alignment Score (WAS). All files share a common identifier "Key" linking records across the dataset.