arXiv AI
LeanPRM: Lean-Grounded Process Supervision for Natural Language Proof Verification
Read the original on arXiv AI →The Flow has not summarised this story yet — read it at arXiv AI.
Related stories
arXiv Machine Learning
Oct 2
LeanPolish: Verified Supervision for Lean Proof Compression
arXiv:2609.38384v1 Announce Type: new Abstract: Verified proof edits offer a natural source of supervision for improving language-model-generated Lean proofs. Yet verification establishes that an edi...
arXiv AI
Jun 12
Pythagoras-Prover: Advancing Efficient Formal Proving via Augmented Lean Formalisation
arXiv:2606. 12594v1 Announce Type: new Abstract: Modern Lean theorem provers achieve strong performance only with substantial training and inference compute, driven in part by scarce verified proof data and the long reasoning traces of formal proof search, making both supervised fine-tuning (SFT) and sampling expensive.
arXiv AI
Sep 12
Magenta: Closing the Loop Between Mathematical Reasoning and Lean Verification
Magenta is a training‑free pipeline that bridges informal natural‑language mathematical problems and formal Lean 4 verification. Given a problem in plain text, it generates an answer, translates it into a Lean 4 statement, and constructs a machine‑checked proof. The system includes a statement judge to ensure the formalisation matches the original problem and an error‑attribution judge to guide corrections, achieving perfect accuracy on olympiad benchmarks and solving all six IMO 2026 problems when combined with K2‑Horizon‑7B.
arXiv AI
Aug 28
FaithSieve: Fine-Grained Evaluation of Math Proofs with Faithful Formal Evidence
FaithSieve is a Lean‑assisted framework that fine‑grains natural‑language mathematical proofs into local reasoning units, extracts typed proof obligations, and verifies them with formal evidence gated by semantic alignment. It introduces two expert‑verified datasets—ProofLoc‑Olympiad and ProofLoc‑University—to benchmark first‑error localization. On these benchmarks, FaithSieve outperforms direct‑judging baselines, achieving 81.43% and 84.5% exact first‑error accuracy respectively.
arXiv Machine Learning
Oct 16
Learning Lean Feedback for Efficient Proof-Candidate Evaluation on GPUs
The paper presents a neural evaluator that predicts Lean 4 theorem‑proving verdicts and compiler feedback, trained on 4.84 million labelled programs. A two‑stage approach—first screening candidates with a validity classifier, then scoring positives with a feedback model—achieves 87.8 % accuracy on a 68,932‑program benchmark and processes 3.5 candidates per second on an H100 GPU, outperforming a 64‑worker Lean baseline. Integrating this evaluator into a hybrid reward system reduces GRPO training steps by about 31 % while maintaining comparable miniF2F pass rates.
arXiv AI
Jul 1
Beyond Compilation: Evaluating Faithful Natural-Language-to-Lean Statement Formalization
arXiv:2606. 31002v1 Announce Type: new Abstract: Theorem-proving benchmarks evaluate proof search against fixed formal statements, but natural-language-to-Lean formalization must generate the formal statement itself.