Preserving Privacy in Large Language Models: A Survey on Current Threats and Solutions