ReuF2F: Benchmarking Coding Agents on Formalized Undergraduate Research Problems
A benchmark of 244 open problems from undergraduate research, formalized in Lean 4, for evaluating coding agents on proving statements or their negations.
Mathematics and logic at Carnegie Mellon University.
I go by Louis and publish as Yiyang Liu.
Proof theory, formalized mathematics, and automated theorem proving.
I study mathematics, with a focus on discrete mathematics and logic, at Carnegie Mellon University. I am interested in the structure of proofs and in using proof assistants to make mathematical arguments precise. I write proofs in Lean.
A benchmark of 244 open problems from undergraduate research, formalized in Lean 4, for evaluating coding agents on proving statements or their negations.
I am working on an axiomatization of classical mechanics.
I also write about mathematics and logic in my notes and longer series.