Photo by Athena Sandrini on Pexels

A new study explores the capability of large language models (LLMs) to deduce the origins of names using a data-driven approach, mirroring how historians and geographers analyze name etymology. The researcher devised a two-stage process to guide the LLM in making informed inferences about name origins. The prompts used in the study, along with the resulting origin predictions, are publicly available for review and scrutiny at https://nameplay.org/blog/educating-name-meaning-guesses-with-data and https://nameplay.org/list/names-with-inferred-origins. The primary goal is to determine if this method produces more accurate and refined insights into name origins compared to existing name databases, potentially reducing inaccuracies and biases. Initial discussions and feedback regarding this research took place on Reddit: https://old.reddit.com/r/artificial/comments/1mcgehl/can_an_llm_make_educated_guesses_about_name/