Can LLMs Replace Humans During Code Chunking?