27B Model Predicts Mathematical Theorem Interestingness via Proof Difficulty
September 22, 2026
A 27B parameter model predicts the intrinsic interestingness of mathematical theorems by analyzing the ratio of proof length to statement length. The method identifies theorem difficulty as a primitive to correlate mathematical conjectures with downstream utility.
HOW THIS AFFECTS YOU
●
researcherThis introduces a scalable way to filter mathematically significant conjectures from noise using proof-to-statement ratios.