# Metamath set.mm: 63,141 Formally Verified Theorems and Axioms in Structured Form (Metamath, phanerozoic extraction)

## Facts

- Provider: Metamath contributors (Norman Megill et al.); extraction by phanerozoic
- License: CC0-1.0 (https://github.com/metamath/set.mm/blob/develop/LICENSE)
- Size: 3 files
- Format: Parquet (in sampled files)
- Stored at: Hugging Face, https://huggingface.co/datasets/phanerozoic/Metamath
- Access: public, readable without an account
- Category: Mathematics
- Source: https://us.metamath.org
- Page: https://scigantic.com/archive/84c723ac-38f5-5570-85c5-45f4bbd21ebb
- Keywords: Mathematics, Parquet, Hugging Face

## About

The Metamath databases of formally verified mathematics (set.mm for ZFC set theory and classical logic, plus hol.mm and the other main libraries in the metamath/set.mm repository), split into one row per declaration. There are 63,141 entries: 59,651 proved theorems ($p) and 3,490 axioms and definitions ($a). The extraction is from upstream commit 160dfc7e. Each row has the statement (the claim without its proof), the verbatim compressed proof, the declaration type, the label (symbolic_name), the sub-library and file, the intra-corpus identifiers it depends on, and the preceding documentation comment. Metamath proofs are checked by small independent verifiers, and set.mm is one of the largest bodies of fully machine-checked mathematics. That makes it a corpus for proof modeling, automated theorem proving, retrieval and dependency analysis. It is one Parquet file; load it with pandas or Hugging Face datasets. The upstream Metamath databases are dedicated to the public domain under CC0 1.0 by their contributors. Credit the Metamath project (us.metamath.org).

## Load the data

```python
from datasets import load_dataset
ds = load_dataset("phanerozoic/Metamath")
print(ds)
```

## Data structure

File types in sampled files: Parquet

Example file: `data/train-00000-of-00001.parquet`

## Open it

- In a hosted notebook with the data mounted: https://scigantic.com/archive/84c723ac-38f5-5570-85c5-45f4bbd21ebb
- Full record as JSON (schema card included): https://api.scigantic.com/api/archives/84c723ac-38f5-5570-85c5-45f4bbd21ebb
- Over MCP: `get_schema_card` with id `84c723ac-38f5-5570-85c5-45f4bbd21ebb` at https://api.scigantic.com/mcp

## Related topics

- [Parquet datasets](https://scigantic.com/topics/parquet)
- [Mathematics datasets](https://scigantic.com/topics/mathematics)
