Guides for building a labeled evaluation set and measuring your heuristic baseline.
| Document | Purpose |
|---|---|
| annotation_guidelines.md | What to label, how to label it, and how to stay consistent across PDFs |
| evaluation_and_baseline_reporting.md | Metrics definitions and how to score prospectus-analyze against gold labels (by hand or with a script later) |
Start with annotation_guidelines.md, then read evaluation_and_baseline_reporting.md once you have labels for at least a few documents.