Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,79 @@
|
|
| 1 |
-
---
|
| 2 |
-
license: mit
|
| 3 |
-
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: mit
|
| 3 |
+
---
|
| 4 |
+
# 📜 ProofWalaDataset
|
| 5 |
+
|
| 6 |
+
The **ProofWalaDataset** is a multilingual dataset of formal theorem proving traces collected from multiple interactive theorem prover (ITP) ecosystems. It provides a structured view of proof steps, goals, hypotheses, and theorem names from diverse mathematical and program verification libraries.
|
| 7 |
+
|
| 8 |
+
This dataset is intended for researchers and practitioners working on:
|
| 9 |
+
- **Automated theorem proving**
|
| 10 |
+
- **Formal code generation**
|
| 11 |
+
- **Machine learning for logic**
|
| 12 |
+
- **Proof step prediction**
|
| 13 |
+
- **Multi-language transfer in formal systems**
|
| 14 |
+
|
| 15 |
+
---
|
| 16 |
+
|
| 17 |
+
## 📂 Dataset Structure
|
| 18 |
+
|
| 19 |
+
The dataset is organized into the following ITP families:
|
| 20 |
+
|
| 21 |
+
- `lean/`
|
| 22 |
+
- `coq/`
|
| 23 |
+
- `GeoCoq/`
|
| 24 |
+
- `math-comp/`
|
| 25 |
+
- `multilingual/` (cross-formal-language hybrid)
|
| 26 |
+
|
| 27 |
+
Each family includes standard splits:
|
| 28 |
+
**`train/`**, **`test/`**, and **`eval/`**, each containing multiple JSON files.
|
| 29 |
+
Each JSON file contains a top-level key: `"training_data"` with a list of proof records.
|
| 30 |
+
|
| 31 |
+
---
|
| 32 |
+
|
| 33 |
+
## 🔍 Each record contains
|
| 34 |
+
|
| 35 |
+
| Field | Description |
|
| 36 |
+
|------------------------------|-------------|
|
| 37 |
+
| `proof_id` | Unique identifier for the proof trace |
|
| 38 |
+
| `goal_description` | Optional natural language description of the proof |
|
| 39 |
+
| `start_goals` | List of starting goals (each with `goal` and `hypotheses`) |
|
| 40 |
+
| `end_goals` | Final goals after applying proof steps |
|
| 41 |
+
| `proof_steps` | List of applied proof tactics (`inv`, `rewrite`, etc.) |
|
| 42 |
+
| `simplified_goals` | Simplified representations of goals (if any) |
|
| 43 |
+
| `all_useful_defns_theorems` | Set of useful definitions or theorems (static analysis) |
|
| 44 |
+
| `addition_state_info` | Optional additional metadata about the proof context |
|
| 45 |
+
| `file_path` | Source file where the proof appears |
|
| 46 |
+
| `project_id` | The ITP project or repository path (e.g., CompCert) |
|
| 47 |
+
| `theorem_name` | Name of the theorem being proved |
|
| 48 |
+
|
| 49 |
+
For convenience, structured fields such as `start_goals[*].goal`, `start_goals[*].hypotheses`, `end_goals[*].goal`, and `end_goals[*].hypotheses` are exposed directly through the Croissant metadata.
|
| 50 |
+
|
| 51 |
+
---
|
| 52 |
+
|
| 53 |
+
## 🧠 Use Cases
|
| 54 |
+
|
| 55 |
+
- Pretraining and finetuning **LLMs for formal verification**
|
| 56 |
+
- Evaluating **proof search** strategies
|
| 57 |
+
- Building **cross-language proof translators**
|
| 58 |
+
- Fine-grained **proof tactic prediction**
|
| 59 |
+
|
| 60 |
+
---
|
| 61 |
+
|
| 62 |
+
## 📊 Format
|
| 63 |
+
|
| 64 |
+
- Data format: **JSON**
|
| 65 |
+
- Schema described via **Croissant metadata** (`croissant.json`)
|
| 66 |
+
- Fully validated using [mlcroissant](https://github.com/mlcommons/croissant)
|
| 67 |
+
|
| 68 |
+
---
|
| 69 |
+
|
| 70 |
+
## 📜 Citation
|
| 71 |
+
|
| 72 |
+
```bibtex
|
| 73 |
+
@misc{thakur2024proofwala,
|
| 74 |
+
title={ProofWala Dataset},
|
| 75 |
+
author={Amitayush Thakur},
|
| 76 |
+
year={2024},
|
| 77 |
+
howpublished={\url{https://huggingface.co/datasets/amitayusht/ProofWalaDataset}}
|
| 78 |
+
}
|
| 79 |
+
|