A template for blueprint-driven formalization projects in Lean.

Guiding Search for Neural Theorem Proving via Proof Progress Prediction

構成的素数構造（6n±1）と除去関数により、Goldbach予想・Bertrand仮説・Catalan予想を共通基盤から再構成。非構成的証明とも形式整合し、論理強度と実行再現性の両立を実現。GitHubにて統合証明パッケージとして公開。 Using the 6n±1 prime structure and composite exclusion, we reconstruct and unify constructive proofs of Goldbach, Bertrand, and Catalan conjectures, demonstrating formal alignment with classical non-constructive methods.