Loop Engineering for Hierarchical Retrieval: Reading a Long Document by Its Table of Contents

Finding specific information in massive documents can be flawed, but a new technique mimics how an expert navigates a table of contents to pinpoint answers. This hierarchical retrieval method avoids overwhelming a language model by first presenting only the top-level chapter titles for a decision. After the model chooses a branch, the system iteratively descends through the document's structure, revealing the next level of subheadings for another choice. This focused, step-by-step process dramatically improves precision and efficiency by zeroing in on the exact section needed without processing irrelevant content.