Archive of benchmark results measuring token usage and retrieval accuracy across file formats for LLM consumption to find the most efficient format. This repository contains benchmark results and raw ...