README.md

---
language:
- en
license: cc-by-sa-4.0
size_categories:
- 1M<n<10M
task_categories:
- text-generation
pretty_name: SEC EDGAR Material Contracts (Exhibit 10)
tags:
- legal
- finance
dataset_info:
  features:
  - name: index_html_url
    dtype: large_string
  - name: index_text_url
    dtype: large_string
  - name: cik
    dtype: large_string
  - name: name
    dtype: large_string
  - name: type
    dtype: large_string
  - name: filing_date
    dtype: large_string
  - name: report_date
    dtype: large_string
  - name: seq
    dtype: large_string
  - name: desc
    dtype: large_string
  - name: doc_type
    dtype: large_string
  - name: size
    dtype: large_string
  - name: filename
    dtype: large_string
  - name: file_url
    dtype: large_string
  - name: file
    dtype: large_string
  - name: extension
    dtype: large_string
  - name: filing_metadata
    dtype: large_string
  - name: file_content
    dtype: string
  splits:
  - name: train
    num_bytes: 151531667922.6878
    num_examples: 1141632
  download_size: 39664203631
  dataset_size: 151531667922.6878
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---

# <span style="color: #2c2c2c">Material Contracts (Exhibit 10) from SEC/EDGAR</span>

*<span style="color: #81b29a">Because sometimes you need 1,141,632 examples of corporate legalese to train your next model</span>* ☕

## Table of Contents
- [Dataset Card Creation Guide](#dataset-card-creation-guide)
  - [Table of Contents](#table-of-contents)
  - [Dataset Description](#dataset-description)
    - [Dataset Summary](#dataset-summary)
    - [Supported Tasks and Leaderboards](#supported-tasks-and-leaderboards)
    - [Languages](#languages)
  - [Dataset Structure](#dataset-structure)
    - [Data Instances](#data-instances)
    - [Data Fields](#data-fields)
    - [Data Splits](#data-splits)
  - [Dataset Creation](#dataset-creation)
    - [Curation Rationale](#curation-rationale)
    - [Source Data](#source-data)
      - [Initial Data Collection and Normalization](#initial-data-collection-and-normalization)
      - [Who are the source language producers?](#who-are-the-source-language-producers)
    - [Annotations](#annotations)
      - [Annotation process](#annotation-process)
      - [Who are the annotators?](#who-are-the-annotators)
    - [Personal and Sensitive Information](#personal-and-sensitive-information)
  - [Considerations for Using the Data](#considerations-for-using-the-data)
    - [Social Impact of Dataset](#social-impact-of-dataset)
    - [Discussion of Biases](#discussion-of-biases)
    - [Other Known Limitations](#other-known-limitations)
  - [Additional Information](#additional-information)
    - [Dataset Curators](#dataset-curators)
    - [Licensing Information](#licensing-information)
    - [Citation Information](#citation-information)
    - [Contributions](#contributions)

## Dataset Description

- **Homepage:** NA
- **Repository:** [Scrapy Crawler](https://github.com/ChenghaoMou/edgar-crawler/tree/main)
- **Paper:** NA
- **Leaderboard:** NA
- **Point of Contact:** mouchenghao at gmail dot com


### <span style="color: #e07a5f">Dataset Summary</span>

Picture this: <span style="color: #e07a5f">**1,141,632 material contracts**</span> (Exhibit 10) painstakingly collected from sec.gov's EDGAR database. We're talking about legal agreements spanning from <span style="color: #81b29a">**1994 to 2025 Q1**</span>, sourced from 10-K, 10-Q, and 8-K filings. Think of Exhibit 10 as the treasure trove where companies hide their most important legal paperwork – employment agreements, merger documents, licensing deals, and all those contracts that make corporate lawyers reach for their third espresso.

This isn't just another dataset – it's <span style="color: #81b29a">**three decades of corporate America's legal DNA**</span>, ready for your next language model to digest.

### <span style="color: #e07a5f">Supported Tasks and Leaderboards</span>

- <span style="color: #2c2c2c; background-color: #f4f3ee; padding: 2px 4px; border-radius: 3px;">`language-modeling`</span> or <span style="color: #2c2c2c; background-color: #f4f3ee; padding: 2px 4px; border-radius: 3px;">`text-generation`</span>: Perfect for building domain-specific models that understand the intricate dance of legal and financial language. Your model will learn to speak fluent "corporate" – complete with all those delightful <span style="color: #81b29a">*whereases*</span> and <span style="color: #81b29a">*heretofores*</span>.

Currently, there's no leaderboard for this dataset, but hey, that's an <span style="color: #e07a5f">**opportunity**</span> waiting for someone with enough coffee and determination!

### <span style="color: #e07a5f">Languages</span>

Primarily <span style="color: #81b29a">**US English**</span>, though you might encounter the occasional foreign phrase when companies get international. It's like <span style="color: #81b29a">*linguistic archaeology*</span> – you never know what you'll dig up.

## <span style="color: #e07a5f">Dataset Structure</span>

### <span style="color: #e07a5f">Data Instances</span>

Check out the data viewer for examples – trust me, it's more interesting than it sounds. Each instance is like a <span style="color: #81b29a">*little corporate story*</span> waiting to be told.

### <span style="color: #e07a5f">Data Fields</span>

Here's what each record contains <span style="color: #81b29a">*(because data dictionaries are love letters to future developers)*</span>:

| Field | Description |
|---|---|
| `index_html_url` | Filing index page *(your breadcrumb trail back to the source)* |
| `index_text_url` | Filing index text page *(for when HTML isn't your thing)* |
| `cik` | Central Index Key from EDGAR *(think of it as a company's social security number)* |
| `name` | Company name *(who's responsible for this legal masterpiece)* |
| `type` | Filing type *(10-K, 10-Q, or 8-K – the holy trinity of SEC filings)* |
| `filing_date` | Filing date |
| `report_date` | Report date |
| `seq` | Sequence number in the filing |
| `desc` | Description provided from the filing *(sometimes helpful, sometimes cryptic)* |
| `doc_type` | Document type (e.g. EX-10) *(the exhibit classification)* |
| `size` | Document size *(bigger isn't always better, but it usually means more billable hours)* |
| `filename` | Document name *(often more creative than you'd expect)* |
| `file_url` | Document page URL *(your direct line to the source)* |
| `file` | GCS file URI *(private, like a good secret)* |
| `extension` | txt, html, pdf, or txt_filing |
| `file_content` | Text content (HTML) or base64 string for binary content (PDF) *(the good stuff)* |

Please be aware, `txt_filing` means the `file_content` is extracted from the text filing file instead of individual files. This happens to most filings before 2000.

### <span style="color: #e07a5f">Data Splits</span>

There's no split by design – everything lives under <span style="color: #2c2c2c; background-color: #f4f3ee; padding: 2px 4px; border-radius: 3px;">`train`</span> because sometimes life is simpler that way. Feel free to create your own splits based on your specific needs <span style="color: #81b29a">*(and caffeine levels)*</span>.

By year:

```
┌──────┬───────┐
│ year ┆ count │
│ ---  ┆ ---   │
│ i32  ┆ u32   │
╞══════╪═══════╡
│ 1994 ┆ 6795  │
│ 1995 ┆ 9948  │
│ 1996 ┆ 19463 │
│ 1997 ┆ 28942 │
│ 1998 ┆ 30261 │
│ 1999 ┆ 30582 │
│ 2000 ┆ 30100 │
│ 2001 ┆ 30372 │
│ 2002 ┆ 32698 │
│ 2003 ┆ 35833 │
│ 2004 ┆ 40178 │
│ 2005 ┆ 54450 │
│ 2006 ┆ 53832 │
│ 2007 ┆ 51678 │
│ 2008 ┆ 52070 │
│ 2009 ┆ 49801 │
│ 2010 ┆ 43433 │
│ 2011 ┆ 42202 │
│ 2012 ┆ 38889 │
│ 2013 ┆ 38346 │
│ 2014 ┆ 38721 │
│ 2015 ┆ 39052 │
│ 2016 ┆ 36937 │
│ 2017 ┆ 34964 │
│ 2018 ┆ 34266 │
│ 2019 ┆ 33453 │
│ 2020 ┆ 35807 │
│ 2021 ┆ 40322 │
│ 2022 ┆ 36202 │
│ 2023 ┆ 35707 │
│ 2024 ┆ 35014 │
│ 2025 ┆ 21314 │
└──────┴───────┘
```

By extension:

```
┌────────────┬────────┐
│ extension  ┆ count  │
│ ---        ┆ ---    │
│ str        ┆ u32    │
╞════════════╪════════╡
│ html       ┆ 854986 │
│ txt        ┆ 142454 │
│ txt_filing ┆ 140820 │
│ pdf        ┆ 3372   │
└────────────┴────────┘
```

## <span style="color: #e07a5f">Dataset Creation</span>

### <span style="color: #e07a5f">Curation Rationale</span>

Why spend <span style="color: #e07a5f">**sleepless nights**</span> building this dataset? Because SEC EDGAR is a <span style="color: #81b29a">**goldmine**</span> of publicly available corporate intelligence, and someone had to do the heavy lifting of making it ML-ready. This collection of exhibit files gives researchers direct access to the contracts and agreements that drive corporate decision-making – no more parsing through entire filings to find the good stuff.

### <span style="color: #e07a5f">Source Data</span>

https://www.sec.gov/ *<span style="color: #81b29a">(the government's gift to data scientists everywhere)</span>*

#### <span style="color: #e07a5f">Initial Data Collection and Normalization</span>

The data collection process was like archaeological excavation, but with more <span style="color: #81b29a">**Python**</span> and less dirt. We crawled through years of filings (10-K, 8-K, and 10-Q), extracting each exhibit individually with complete metadata. It's the kind of methodical, detail-oriented work that requires <span style="color: #e07a5f">multiple monitors</span> and a steady supply of debugging fuel.

Each document was downloaded with <span style="color: #81b29a">*surgical precision*</span>, preserving all metadata for maximum research utility. Think of it as digital preservation, but for corporate paperwork.

#### <span style="color: #e07a5f">Who are the source language producers?</span>

The unsung heroes behind this dataset are the army of <span style="color: #81b29a">**finance professionals, lawyers, and corporate executives**</span> who craft these documents. The SEC requires public companies, insiders, and broker-dealers to file these periodic statements, creating a continuous stream of corporate communication that investors and researchers rely on for informed decision-making.

These documents represent the <span style="color: #81b29a">*collective voice of corporate America*</span> – sometimes eloquent, sometimes bureaucratic, always fascinating from a linguistic perspective.

### Annotations

#### Annotation process

No additional annotations were added – we kept it pure and unadulterated, just the way the SEC intended.

#### Who are the annotators?

The original authors and their legal teams. Every comma placement was probably billable.

### Personal and Sensitive Information

Fair warning: this dataset might contain PII (names, emails, job titles, companies) that's already in the public domain. It comes with the territory when dealing with public filings – executives' names and contact information are part of the transparency requirements.

## <span style="color: #e07a5f">Considerations for Using the Data</span>

### <span style="color: #e07a5f">Social Impact of Dataset</span>

This dataset could be the catalyst for <span style="color: #81b29a">**better legal and finance language modeling capabilities**</span>. Imagine AI systems that can parse complex contracts, identify key terms, or help democratize access to legal understanding. Of course, with great power comes great responsibility – <span style="color: #e07a5f">*use it wisely*</span>.

### <span style="color: #e07a5f">Discussion of Biases</span>

Given the source and language, this dataset has a distinctly <span style="color: #81b29a">**US-centric flavor**</span> with a heavy dose of corporate legalese. It reflects the linguistic patterns and legal frameworks of American business culture, which might not translate well to other jurisdictions or informal contexts. Your model might end up sounding like it went to law school <span style="color: #81b29a">*(and accumulated the corresponding debt)*</span>.

### <span style="color: #e07a5f">Other Known Limitations</span>

Like any dataset scraped from the wild, this one has its quirks. Filing content can follow different formats based on their filing softwares and time, and legal documents have their own special way of torturing the English language. Approach with realistic expectations and a <span style="color: #e07a5f">**good debugger**</span>.

## <span style="color: #e07a5f">Additional Information</span>

### <span style="color: #e07a5f">Dataset Curators</span>

@chenghao *<span style="color: #81b29a">(the sleep-deprived soul who made this possible)</span>*

### <span style="color: #e07a5f">Licensing Information</span>

[Attribution-ShareAlike 4.0 International](https://creativecommons.org/licenses/by-sa/4.0/) *<span style="color: #81b29a">(share and share alike, just the way open source should be)</span>*

### <span style="color: #e07a5f">Citation Information</span>

NA *<span style="color: #81b29a">(but feel free to give credit where credit is due)</span>*

### <span style="color: #e07a5f">Contributions</span>

This dataset exists thanks to countless hours of <span style="color: #e07a5f">**debugging**</span>, data wrangling, and the unwavering belief that good datasets make the world a better place. Special thanks to the SEC for maintaining EDGAR and making corporate transparency possible.

---

*<span style="color: #81b29a">Happy modeling! May your training runs be swift and your validation losses ever-decreasing.</span>* 🚀

**<span style="color: #e07a5f">sleepless</span><span style="color: #2c2c2c">@debugging:</span><span style="color: #81b29a">/datasets</span><span style="color: #2c2c2c">$</span> <span style="color: #e07a5f">█</span>**