Evaluation of Large Language Models' Software Vulnerability Detection Performance Across Common Weakness Enumeration Categories