Pith. sign in

REVIEW 9 cited by

LEGO-Prover: Neural Theorem Proving with Growing Libraries

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2310.00656 v3 pith:A6OWPH33 submitted 2023-10-01 cs.AI

LEGO-Prover: Neural Theorem Proving with Growing Libraries

classification cs.AI
keywords provingskillslibrarylego-provertheoremtheoremsduringgrowing
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
read the original abstract

Despite the success of large language models (LLMs), the task of theorem proving still remains one of the hardest reasoning tasks that is far from being fully solved. Prior methods using language models have demonstrated promising results, but they still struggle to prove even middle school level theorems. One common limitation of these methods is that they assume a fixed theorem library during the whole theorem proving process. However, as we all know, creating new useful theorems or even new theories is not only helpful but crucial and necessary for advancing mathematics and proving harder and deeper results. In this work, we present LEGO-Prover, which employs a growing skill library containing verified lemmas as skills to augment the capability of LLMs used in theorem proving. By constructing the proof modularly, LEGO-Prover enables LLMs to utilize existing skills retrieved from the library and to create new skills during the proving process. These skills are further evolved (by prompting an LLM) to enrich the library on another scale. Modular and reusable skills are constantly added to the library to enable tackling increasingly intricate mathematical problems. Moreover, the learned library further bridges the gap between human proofs and formal proofs by making it easier to impute missing steps. LEGO-Prover advances the state-of-the-art pass rate on miniF2F-valid (48.0% to 57.0%) and miniF2F-test (45.5% to 47.1%). During the proving process, LEGO-Prover also manages to generate over 20,000 skills (theorems/lemmas) and adds them to the growing library. Our ablation study indicates that these newly added skills are indeed helpful for proving theorems, resulting in an improvement from a success rate of 47.1% to 50.4%. We also release our code and all the generated skills.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 9 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Verifiable Auto-Formalization of Mathematics Using a Relaxed Natural Formal Language

    cs.LO 2026-06 unverdicted novelty 7.0

    Introduces Relaxed NFL intermediate language for LLM-based auto-formalization, with rule-plus-LLM elaboration to Core NFL and tactic-language discharge of verification conditions.

  2. Formalizing Mathematics at Scale

    cs.AI 2026-05 accept novelty 7.0

    A multi-agent framework called AutoformBot autoformalized 26 textbooks spanning analysis, algebra, topology, combinatorics and probability into a verified Lean 4 library of 45k declarations, demonstrating scalable for...

  3. Keep the Proof State Live: Snapshotting for Efficient Tactic Search in Lean 4

    cs.LO 2026-05 unverdicted novelty 7.0

    Proof-state snapshotting in Lean 4 delivers 5.6-50x wall-time speedups (average 14x) on 48 miniF2F-v2 problems by reusing elaborated states and avoiding repeated import loading and theorem-body elaboration.

  4. Less Effort, Shorter Proofs: Reinforcement Learning for Security Protocol Analysis in Tamarin

    cs.CR 2026-05 unverdicted novelty 7.0

    An RL-guided MCTS proof search for Tamarin finds more and shorter proofs than standard search across 16 protocol models.

  5. LiveFMBench: Unveiling the Power and Limits of Agentic Workflows in Specification Generation

    cs.SE 2026-05 conditional novelty 7.0

    LiveFMBench shows that direct LLM prompting for C program formal specs overestimates accuracy by ~20% due to unfaithful behaviors like deceiving provers, while agentic workflows help under low sampling but overall per...

  6. Kimina-Prover Preview: Towards Large Formal Reasoning Models with Reinforcement Learning

    cs.AI 2025-04 unverdicted novelty 6.0

    Kimina-Prover achieves 80.7% pass@8192 on miniF2F for Lean 4 theorem proving via RL training and a formal reasoning pattern that emulates iterative human-style proof refinement.

  7. IsabeLLM: Automated Theorem Proving Applied to Formally Verifying Consensus

    cs.AI 2026-06 unverdicted novelty 4.0

    Improved IsabeLLM adds RAG, error tracing, and Isabelle compatibility updates and is tested on verifying Bitcoin Proof of Work consensus.

  8. Rethinking Wireless Communications through Formal Mathematical AI Reasoning

    eess.SP 2026-04 unverdicted novelty 4.0

    Proposes a three-layer framework using formal AI reasoning for verification, derivation, and discovery in wireless communications theory.

  9. A Survey of Self-Evolving Agents: What, When, How, and Where to Evolve on the Path to Artificial Super Intelligence

    cs.AI 2025-07 accept novelty 4.0

    The paper delivers the first systematic review of self-evolving agents, structured around what components evolve, when adaptation occurs, and how it is implemented.