Language Models Exhibit Inconsistent Biases Towards Algorithmic Agents and Human Experts

arXiv:2602.22070v1 Announce Type: new Abstract: Large language models are increasingly used in decision-making tasks that require them to process information from a variety of sources, …