If the LLM was trained on a corpus of pairs of equivalent Greek/Hebrew and English texts, its training data would almost certainly have included the Bible.