SpreadsheetLLM: Encoding Spreadsheets for Large Language Models

I'd really like to try this out! Or try out something built based on this paper.