Search papers, labs, and topics across Lattice.
This paper introduces Visored, a dependent-type-based prover tailored to align with the mathematical writing style of LLMs and humans, enhancing the usability of formal proof systems like Lean and Rocq. By integrating a natural language surface with a rule-driven automation layer, Visored streamlines the proof process, allowing users to generate checked Lean files from accepted proofs. Initial experiments indicate that LLMs can effectively utilize Visored on the miniF2F benchmark, even without specific training data for the prover.
LLMs can seamlessly generate formal proofs with a new system that mimics natural mathematical language, bridging the gap between human intuition and formal verification.
We present a dependent-type-based prover designed around the way LLMs (and humans) tend to write mathematics, complementing existing systems such as Lean and Rocq. Its core design choices are a surface that imitates mathematical natural language and a rule-driven automation layer that closes the routine steps a textbook would omit, so that an accepted proof can be re-emitted as a checked Lean file. Early experiments suggest that, even without any prover-specific training data, LLMs can learn to use it effectively on the miniF2F benchmark. Lean output excerpts: https://github.com/xiyuzhai-husky-lang/visored/
UW