README.md

---
license: apache-2.0
pretty_name: ProofWriter (The Source)
tags:
- logical-reasoning
- proof-generation
configs:
- config_name: OWA-depth-0
  data_files:
  - split: train
    path: OWA-depth-0/train.parquet
  - split: dev
    path: OWA-depth-0/dev.parquet
  - split: test
    path: OWA-depth-0/test.parquet
- config_name: OWA-depth-1
  data_files:
  - split: train
    path: OWA-depth-1/train.parquet
  - split: dev
    path: OWA-depth-1/dev.parquet
  - split: test
    path: OWA-depth-1/test.parquet
- config_name: OWA-depth-2
  data_files:
  - split: train
    path: OWA-depth-2/train.parquet
  - split: dev
    path: OWA-depth-2/dev.parquet
  - split: test
    path: OWA-depth-2/test.parquet
- config_name: OWA-depth-3
  data_files:
  - split: train
    path: OWA-depth-3/train.parquet
  - split: dev
    path: OWA-depth-3/dev.parquet
  - split: test
    path: OWA-depth-3/test.parquet
- config_name: OWA-depth-5
  data_files:
  - split: train
    path: OWA-depth-5/train.parquet
  - split: dev
    path: OWA-depth-5/dev.parquet
  - split: test
    path: OWA-depth-5/test.parquet
- config_name: CWA-depth-0
  data_files:
  - split: train
    path: CWA-depth-0/train.parquet
  - split: dev
    path: CWA-depth-0/dev.parquet
  - split: test
    path: CWA-depth-0/test.parquet
- config_name: CWA-depth-1
  data_files:
  - split: train
    path: CWA-depth-1/train.parquet
  - split: dev
    path: CWA-depth-1/dev.parquet
  - split: test
    path: CWA-depth-1/test.parquet
- config_name: CWA-depth-2
  data_files:
  - split: train
    path: CWA-depth-2/train.parquet
  - split: dev
    path: CWA-depth-2/dev.parquet
  - split: test
    path: CWA-depth-2/test.parquet
- config_name: CWA-depth-3
  data_files:
  - split: train
    path: CWA-depth-3/train.parquet
  - split: dev
    path: CWA-depth-3/dev.parquet
  - split: test
    path: CWA-depth-3/test.parquet
- config_name: CWA-depth-5
  data_files:
  - split: train
    path: CWA-depth-5/train.parquet
  - split: dev
    path: CWA-depth-5/dev.parquet
  - split: test
    path: CWA-depth-5/test.parquet
---

# ProofWriter (The Source)

An **unmodified copy** of AI2's ProofWriter dataset (release `V2020.12.3`), re-hosted as `datasets`
configs for convenient loading. The records are faithful to the upstream release — the id-keyed JSON
is preserved as-is; typing and reasoning-graph extraction happen in later stages.

Each config is a `{world}-depth-{n}` shelf of the synthetic core (`OWA`/`CWA` × depths
0/1/2/3/5), split `train`/`dev`/`test` (`dev` kept as the corpus names it).

- **Source:** <https://aristo-data-public.s3.amazonaws.com/proofwriter/proofwriter-dataset-V2020.12.3.zip>
- **License:** `apache-2.0` (inherited from `allenai/ruletaker`; the release ships no LICENSE file).
- Not affiliated with or endorsed by AI2; bundled here with attribution.

## Citation

ProofWriter — and the RuleTaker corpora it builds on — are AI2's work. Please cite them:

```bibtex
@inproceedings{tafjord-etal-2021-proofwriter,
    title = "{P}roof{W}riter: Generating Implications, Proofs, and Abductive
             Statements over Natural Language",
    author = "Tafjord, Oyvind and Dalvi, Bhavana and Clark, Peter",
    booktitle = "Findings of the Association for Computational Linguistics: ACL-IJCNLP 2021",
    year = "2021",
    eprint = "2012.13048",
    archivePrefix = "arXiv",
}
```

```bibtex
@inproceedings{clark-etal-2020-transformers,
    title = "Transformers as Soft Reasoners over Language",
    author = "Clark, Peter and Tafjord, Oyvind and Richardson, Kyle",
    booktitle = "Proceedings of the 29th International Joint Conference on
                 Artificial Intelligence (IJCAI-20)",
    year = "2020",
    eprint = "2002.05867",
    archivePrefix = "arXiv",
}
```