---
license: cc-by-4.0
language:
- en
task_categories:
- text-classification
- text-retrieval
- question-answering
tags:
- legal
- court-records
- trial-exhibits
- musk-v-altman
- openai
pretty_name: Musk v. Altman Trial Exhibits
size_categories:
- n<1K
---

# Musk v. Altman — Trial Exhibits

A structured corpus of trial exhibits (PX / DX) admitted in *Elon Musk, et al.
v. Samuel Altman, et al.*, with transcribed text and editorial commentary.
Each row is one exhibit (email thread, Slack thread, text-message extraction,
memo, deck, etc.) with its YAML metadata plus a Markdown body containing the
exhibit's transcribed content and our notes.

## Provenance

- **Source PDFs** are public court records released under Pretrial Order No.
  1, Dkt. 446. They are linked (not redistributed) via `pdf_url`.
- **Transcribed text and editorial commentary** in `body_markdown` were
  produced by the MTS editorial team.

## License

Released under **CC-BY-4.0**. You may redistribute, remix, and build on this
data — including commercially — provided you credit "MTS — musk-v-altman
trial coverage" and link back to the source. The underlying court records
themselves are public domain (U.S. government works / public filings); the
license applies to our transcription and commentary layer.

## Schema

| field                | type    | notes                                                |
|----------------------|---------|------------------------------------------------------|
| `exhibit_id`         | string  | e.g. `"DX-1017"`, `"PX-22"`                          |
| `exhibit`            | string  | display label (e.g. `"DX 1017"`)                     |
| `party`              | string  | `"Defense"` or `"Plaintiff"`                         |
| `type`               | string  | `"Email"`, `"Slack messages"`, `"Text messages"`, ... |
| `admitted_trial_day` | string  | e.g. `"Day 8 (May 6, 2026)"`                         |
| `uploaded_box_pt`    | string  | upload timestamp (Pacific)                           |
| `uploader`           | string  | filing firm                                          |
| `pages`              | int     | PDF page count                                       |
| `size_bytes`         | int     | PDF size                                             |
| `source_pdf`         | string  | filename                                             |
| `pdf_url`            | string  | public link to the source PDF                        |
| `body_markdown`      | string  | transcribed content + commentary (Markdown)          |

Some rows carry additional frontmatter fields — treat the schema as
open-ended and key off `exhibit_id`.

## Loading

```python
from datasets import load_dataset
ds = load_dataset("<your-username>/musk-v-altman-exhibits", split="train")
print(ds[0]["exhibit_id"], ds[0]["type"])
```

## Known limitations

- Some text-message and Slack exhibits contain redactions (`[REDACTED]`).
- OCR artifacts may persist in older filings.
- Editorial commentary reflects best-effort interpretation, not legal advice.

## Citation

> MTS, *Musk v. Altman Trial Exhibits*, 2026. CC-BY-4.0.