---
language:
- en
license: cc0-1.0
task_categories:
- document-question-answering
tags:
- document-splitting
- document-understanding
- tax-forms
- irs
- pdf
- benchmark
pretty_name: PolitTax Document Splitting Benchmark
size_categories:
- n<1K
---

# PolitTax: Document Splitting Benchmark

A benchmark for evaluating document splitting (boundary detection and classification) on U.S. presidential tax returns.

## Overview

| Metric | Value |
|--------|-------|
| PDF files | 30 |
| Total pages | 3,311 |
| Annotated sub-documents | 1,152 |
| Classification types used | 40 |
| Subjects | Biden, Harris, Obama, Trump |
| Tax years | 2009–2023 |

## Task

Given a multi-page PDF containing one or more IRS tax forms, schedules, and supplementary documents, the task is to:

1. **Segment** the document into sub-documents by identifying start and end page boundaries
2. **Classify** each sub-document by its form type (e.g., "Form 1040", "Schedule A (Form 1040)", "supplement")

## Sub-document Type Distribution (top 10)

| Type | Count |
|------|-------|
| supplement | 573 |
| misc_form | 240 |
| Schedule C (Form 1040) | 41 |
| Form 4562 | 40 |
| Schedule K-1 (Form 1065) | 33 |
| Schedule E (Form 1040) | 14 |
| Form 1065 | 14 |
| Schedule M-3 (Form 1120/1065) | 14 |
| Form 8949 | 13 |
| Form 4797 | 13 |

## Files

- **`annotations.jsonl`** — One JSON object per PDF with ground-truth sub-document boundaries
- **`metadata.csv`** — File-level metadata (president, tax year, page count, etc.)
- **`taxonomy.json`** — Full classification schema with 40+ type definitions and splitting instructions
- **`pdfs/`** — Original PDF files organized by subject

## Annotation Format

Each line in `annotations.jsonl` contains:

```json
{
  "file_name": "obama_2009_complete_return.pdf",
  "page_count": 59,
  "sub_documents": [
    {
      "type": "Form 1040",
      "start_page": 1,
      "end_page": 2,
      "identifier": "Barack H. & Michelle L. Obama"
    }
  ]
}
```

- **`type`**: The IRS form or document type. See `taxonomy.json` for the full list with descriptions.
- **`start_page`** / **`end_page`**: 1-indexed, inclusive page range.
- **`identifier`**: A human-readable label distinguishing instances of the same form type (e.g., taxpayer name, entity name).

## Source

All PDFs are publicly released U.S. government records sourced from:
- [Obama White House Archives](https://obamawhitehouse.archives.gov/blog/2010/04/15/president-obama-and-vice-president-biden-s-tax-returns)
- [The American Presidency Project](https://www.presidency.ucsb.edu/documents/white-house-press-release-the-president-and-vice-president-release-their-2023-tax-returns)
- [Tax Notes](https://www.taxnotes.com/presidential-tax-returns)

## License

The source documents are U.S. government works in the public domain. The annotations in this dataset are released under [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/).