BiasKG: Adversarial Knowledge Graphs to Induce Bias in Large Language Models