LemmaHead: RAG Assisted Proof Generation Using Large Language Models

Tianbo Yang [[email protected],edu], Mingqi Yan [[email protected]], Hongyi Zhao [[email protected]], Tianshuo Yang [[email protected]] · 2025

Read LemmaHead: RAG Assisted Proof Generation Using Large Language Models by Tianbo Yang [[email protected],edu], Mingqi Yan [[email protected]], Hongyi Zhao [[email protected]], Tianshuo Yang [[email protected]] (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2501-15797.txt.

Description

Tianbo Yang and co-authors pair a language model with a textbook memory in a 2025 project to help it write math proofs. Models trained to mimic proof style still stumble on logic. Their system, LemmaHead, looks up relevant passages from published texts and feeds them to the model before it answers.

The task they target is automated theorem proving: given a claim, produce a proof in the Lean formal language. Retrieval gives the model near-at-hand facts and lemma shapes instead of pure recall. The project tests whether borrowed context, not just more training, can lift formal reasoning in models.

Catalog record

CREATORTianbo Yang [[email protected],edu], Mingqi Yan [[email protected]], Hongyi Zhao [[email protected]], Tianshuo Yang [[email protected]]
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE17 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2501-15797

Browse the full library · About the Tower of Babel