Political Alignment in Large Language Models: A Multidimensional Audit of Psychometric Identity and Behavioral Bias

Adib Sakhawat, Tahsin Islam, Takia Farhin +3

As large language models (LLMs) are increasingly deployed, understanding how they express political positioning is important for evaluating alignment and downstream effects. We aud…