Dataset for "Uncovering Ableism in Large Language Models' Responses: Hybrid Sentiment and Thematic Analysis Approach for Disability Representation"

Description

This dataset contains 360 responses generated by three large language models (ChatGPT, Gemini, and Microsoft Copilot) using a factorial prompt design varying by disability type, socioeconomic class, language, and setting context. The responses served as the primary corpus for text preprocessing, sentiment analysis, and topic modeling conducted in Orange3, as well as qualitative thematic analysis and consensus-based human ableism scoring conducted by three analysts.

Institutions

• Airlangga University

  East Java, Surabaya

Categories

Licence