Generative Language Modeling for Automated Theorem Proving

Open paper

Benchmark	Model	Rank	Results
automated-theorem-proving-on-metamath-setmm	GPT-f	–	Percentage correct: 56.2