Lean 4 Theorem Repository for Formal Verification

A collection of 29,000 theorems compiled from over 100 Lean 4 repositories. It was created by InternLM to support the development of theorem provers, including the fine-tuned 7B model InternLM2-Step-Prover.

Use Cases

• Fine-tune language models on the 29,000 theorems for formal verification tasks.
• Benchmark theorem prover performance using the compiled Lean 4 repository data.
• Analyze proof structures and patterns across the collection of theorems.

Strengths

• Contains 29,000 theorems, providing a substantial corpus for training.
• Compiled from over 100 distinct Lean 4 repositories, offering diverse sources.
• Associated with a 7B parameter model achieving state-of-the-art results on benchmarks like MiniF2F-test.

Limitations

• The specific columns, file formats, and dataset size are not detailed in the input.
• The dataset is specialized for Lean 4 and formal verification, limiting general applicability.
• The raw description is promotional, focusing on model performance rather than dataset structure.

Provenance

Source

Compiled from over 100 Lean 4 repositories on GitHub.

Collection Method

Compilation of GitHub repositories containing Lean 4 code.

Freshness

Last updated on 2024-07-25.