Papers/2609.35790
🧪 Test?View on arXiv

Sage: Formalization with Semantic Correction

Not provided in the abstract

semantic correctionformalizationneural theorem provingmathematical fidelity
2609.35790
Builder Relevance
80%
1h ago

Abstract

Sage introduces a framework that enhances the formalization of mathematical statements by addressing the limitations of existing neural theorem provers.

Reality Card

Core Claim

Sage reduces answer leakage from 70.9% to 2.7% while achieving 73.3% pass@4 joint compilation and semantic fidelity on the Omni-MATH dataset.

Method / Result

Achieved 87.4% pass@4 verified fidelity on unformalized International Mathematical Olympiad problems.

Limitations

The paper does not specify the authors or provide extensive details on reproducibility concerns.

Paper to code

Verified implementation resources so builders can test the paper’s claims instead of stopping at the abstract.

No verified implementation link has been attached yet. AIBuzzHub will keep this panel separate from unverified search results.
← Back to all papers