In theorem proving, the task of selecting useful premises from a large library to unlock the proof of a given conjecture is crucially important. This presents a challenge for all theorem provers, especially the ones based on language models, due to their relative inability to reason over huge volumes of premises in text form. This paper introduces Thor, a framework integrating language models and automated theorem provers to overcome this difficulty. In Thor, a class of methods called hammers that leverage the power of automated theorem provers are used for premise selection, while all other tasks are designated to language models. Thor increases a language model's success rate on the PISA dataset from 39% to 57%, while solving 8.2% of problems neither language models nor automated theorem provers are able to solve on their own. Furthermore, with a significantly smaller computational budget, Thor can achieve a success rate on the MiniF2F dataset that is on par with the best existing methods. Thor can be instantiated for the majority of popular interactive theorem provers via a straightforward protocol we provide.Preprint. Under review.

Autoformalization is the process of automatically translating from natural language mathematics to formal specifications and proofs. A successful autoformalization system could advance the fields of formal verification, program synthesis, and artificial intelligence. While the long-term goal of autoformalization seemed elusive for a long time, we show large language models provide new prospects towards this goal. We make the surprising observation that LLMs can correctly translate a significant portion (25.3%) of mathematical competition problems perfectly to formal specifications in Isabelle/HOL. We demonstrate the usefulness of this process by improving a previously introduced neural theorem prover via training on these autoformalized theorems. Our methodology results in a new state-of-the-art result on the MiniF2F theorem proving benchmark, improving the proof rate from 29.6% to 35.2%.Recent advances in large language models [7,9] showed promising capabilities of understanding formal languages [8,32]. However, the existing successes are limited to formal languages where there exists a large body of corpus on the web (e.g., Python language). Formal mathematics data is very scarce. For example, one of the largest formal mathematics libraries, the Archive of Formal Proofs, is only 180MB in size, that is less than 0.18% of the training data for the large language model Codex [8]. Moreover, unlike in the case of commonly used programming languages, where natural language docstrings are broadly available, there is almost zero aligned data between natural language and formal mathematics. Therefore, it is unclear the recent successes can directly contribute to the development of autoformalization.In this work, we explore the prospects of autoformalization with large language models. To our surprise, we find that large language models already have a decent capability of formalizing natural † Correspondence to Yuhuai Wu

No abstract