研究人员正在开发新的方法和基准来提高大语言模型（LLMs）的形式数学推理能力。一种名为Diffusion-Proof的方法利用扩散大语言模型（dLLMs）进行定理证明，在ProofNet-Test和MiniF2F-Test等基准测试中表现优于自回归模型，甚至解决了领先模型无法解决的国际数学奥林匹克问题。另一项开发Visored提供了一个旨在通过模仿自然语言和自动化常规步骤来处理大语言模型生成数学的证明器。此外，Mask-Proof引入了一个用于从真实证明中创建自动可检查的掩码步骤任务的管道，以评估大语言模型在步骤级别的推理能力，而MA-ProofBench则提供了一个专门针对数学分析的新基准，突显了当前大语言模型在该复杂领域中的局限性。
AI
arXiv:2606.19399v1 Announce Type: cross Abstract: LLM-based formal provers often collapse rich verifier signals (syntax errors, type mismatches, partial goal progress) into a binary pass/fail bit. We present VERITAS, a zero-shot framework that routes every verifier signal back in…
arXiv cs.LG
TIER_1English(EN)·Slim Barkallah, Luke Bailey, Kaiyue Wen, Mohammed Abouzaid, Tengyu Ma·
arXiv:2605.20531v2 Announce Type: replace-cross Abstract: Reliable verification of proofs remains a bottleneck for training and evaluating AI systems on hard mathematical reasoning. Fully formal proofs, in languages like Lean, are easy to verify because they are unambiguous and m…
arXiv:2606.19315v1 Announce Type: new Abstract: Enhancing the formal math reasoning capabilities of Large Language Models (LLMs) has become a key focus in both mathematical and computer science communities in recent years. While significant progress has been made in using state-o…
Enhancing the formal math reasoning capabilities of Large Language Models (LLMs) has become a key focus in both mathematical and computer science communities in recent years. While significant progress has been made in using state-of-the-art Auto-Regressive (AR) LLMs for formal t…
Enhancing the formal math reasoning capabilities of Large Language Models (LLMs) has become a key focus in both mathematical and computer science communities in recent years. While significant progress has been made in using state-of-the-art Auto-Regressive (AR) LLMs for formal t…
arXiv cs.AI
TIER_1English(EN)·Xiyu Zhai, Xinyi Chen, Yiping Wang, Runlong Zhou, Liao Zhang, Simon S. Du·
arXiv:2606.17581v1 Announce Type: cross Abstract: 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 n…
arXiv:2506.17104v2 Announce Type: replace Abstract: Large language models (LLMs) have shown promising first-order logic (FOL) reasoning capabilities with applications in various areas. However, their effectiveness in complex mathematical reasoning involving multi-step FOL deducti…
arXiv:2606.15258v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly capable of mathematical problem solving and can even assist with research-level proofs, yet we still lack a scalable and reproducible way to measure step-level reasoning in long proofs a…
arXiv:2606.13782v1 Announce Type: new Abstract: Large Language Models (LLMs) have made notable progress in automated theorem proving, yet existing formal benchmarks remain limited in both mathematical coverage and difficulty. Most are concentrated in areas that are easier to form…
