Biases in Large Language Models: Origins, Inventory and Discussion