Navigating the digital spectrum: Assessing political bias, stability, and downstream fairness in Large Language Models
A robust Political Compass Test evaluation framework is introduced that samples 300 configurations across an eight-dimensional perturbation space varying language, framing, instructions, answer format, option order, and persona wording, and results show most models lean Libertarian-Left on average and larger models sho...
Luka Debevc, Nishan Chatterjee, Antoine Doucet et al.
· 0 citations