An Evaluation Benchmark for Autoformalization in Lean4

TL;DR

This paper introduces a new benchmark for evaluating the autoformalization capabilities of large language models in Lean4, revealing current limitations and guiding future improvements in mathematical language understanding.

Contribution

It presents the first comprehensive benchmark for LLM autoformalization in Lean4 and evaluates state-of-the-art models, highlighting areas for future development.

Findings

LLMs show limitations in complex mathematical autoformalization

Current models perform well on simple tasks but struggle with advanced topics

Benchmark sets the stage for future research in LLM autoformalization

Abstract

Large Language Models (LLMs) hold the potential to revolutionize autoformalization. The introduction of Lean4, a mathematical programming language, presents an unprecedented opportunity to rigorously assess the autoformalization capabilities of LLMs. This paper introduces a novel evaluation benchmark designed for Lean4, applying it to test the abilities of state-of-the-art LLMs, including GPT-3.5, GPT-4, and Gemini Pro. Our comprehensive analysis reveals that, despite recent advancements, these LLMs still exhibit limitations in autoformalization, particularly in more complex areas of mathematics. These findings underscore the need for further development in LLMs to fully harness their potential in scientific research and development. This study not only benchmarks current LLM capabilities but also sets the stage for future enhancements in autoformalization.