Reproduce “LLMs Cannot Be Neutral: Gender Sensitivity Across Candidates, Evaluators, and Models” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Danial Amin, Joni Salminen, Bernard J. Jansen

The authors test whether changing assigned gender identities alters AI judgments of unchanged professional narratives in a hypothetical city-council vote. They vary candidate and evaluator identities independently across eight models. Explicit identity substitutions reversed 1.18% of decisions for four decision-focused models versus 3.38% for four general-purpose language models; a model-level test was not significant, limiting category-wide conclusions.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.