Explanations, Prompts, and Formalizations: Arguments for New Norms in LLM-Enabled Mathematical Research
Machine Learning
Summary
The authors discuss how recent use of large language models (LLMs) has helped solve math problems, but current publishing rules don't require sharing the exact prompts or software setups used. They believe it's important to share these details and to formalize results so machines can check them. They also say that since LLM-generated results can be hard to understand, human mathematicians should create clear, intuitive explanations.
large language modelsmathematical conjecturesprompt disclosuresoftware setupformal verificationmachine verificationintuitive explanationsmathematical publishing norms
Authors
Axel Boldt
Abstract
As several mathematical conjectures have recently been settled using large language models (LLMs), the mathematical community has formulated norms and recommendations regarding the publishing of such results. These norms do not cover the disclosure of the prompts and precise software setup used to obtain those results, nor do they require that results be formalized in a manner that allows for machine verification. I argue that both of these are essential. In addition, since LLM-obtained results may be hard to understand, human authors have the responsibility to invent intuitive explanations.