Testifying to their utility to accurately represent abstractions, completion and extrapolation tasks on integer sequences are a frequent part of general human intelligence and aptitude testing ([42, 31]).

Integer sequences are of central importance to the modeling of concepts admitting complete finitary descriptions. We introduce a novel view on the learning of such concepts and lay down a set of benchmarking tasks aimed at conceptual understanding by machine learning models. These tasks indirectly assess model ability to abstract, and challenge them to reason both interpolatively and extrapolatively from the knowledge gained by observing representative examples. To further aid research in knowledge representation and reasoning, we present FACT, the Finitary Abstraction Comprehension Toolkit.

This paper presents a novel benchmark where the large language model (LLM) must write code that computes integer sequences from the Online Encyclopedia of Integer Sequences (OEIS), a widely-used resource for mathematical sequences. The benchmark is designed to evaluate both the correctness of the generated code and its computational efficiency. Our benchmark reveals that the o1 series of models outperform other frontier models from OpenAI, Anthropic, Meta, and Google in accuracy and cheating rates across both easy and hard integer sequences. In order to ensure models do not exploit memorized sequence values, we introduce an automated cheating detection mechanism that flags the use of lookup tables and validated this automation against human cheating evaluations. This benchmark provides a meaningful challenge for current LLMs, offering insights into their mathematical reasoning and code writing capabilities, which can guide future research directions and model development in mathematical reasoning and code synthesis.

Integer sequences are of central importance to the modeling of concepts admitting complete finitary descriptions. We introduce a novel view on the learning of such concepts and lay down a set of benchmarking tasks aimed at conceptual understanding by machine learning models. These tasks indirectly assess model ability to abstract, and challenge them to reason both interpolatively and extrapolatively from the knowledge gained by observing representative examples. To further aid research in knowledge representation and reasoning, we present FACT, the Finitary Abstraction Comprehension Toolkit.

Large language models (LLMs) that do not give consistent answers across contexts are problematic when used for tasks with expectations of consistency, e.g., question-answering, explanations, etc. Our work presents an evaluation benchmark for self-consistency in cases of under-specification where two or more answers can be correct. We conduct a series of behavioral experiments on the OpenAI model suite using an ambiguous integer sequence completion task. We find that average consistency ranges from 67\% to 82\%, far higher than would be predicted if a model's consistency was random, and increases as model capability improves. Furthermore, we show that models tend to maintain self-consistency across a series of robustness checks, including prompting speaker changes and sequence length changes. These results suggest that self-consistency arises as an emergent capability without specifically training for it. Despite this, we find that models are uncalibrated when judging their own consistency, with models displaying both over- and under-confidence. We also propose a nonparametric test for determining from token output distribution whether a model assigns non-trivial probability to alternative answers. Using this test, we find that despite increases in self-consistency, models usually place significant weight on alternative, inconsistent answers. This distribution of probability mass provides evidence that even highly self-consistent models internally compute multiple possible responses.

The search for abstract patterns is one of the principal occupations of mathematicians. The discovery of similar patterns across different mathematical fields often leads to surprising connections. Probably the most famous example of such an unexpected connection in mathematics is the Taniyama-Shimura conjecture proved in 2001 [2]. It relates elliptic curves over the field of rational numbers with a special kind of complex analytical functions known as modular forms. This conjecture became especially famous because a restricted version of it implies Fermat's last theorem [16]. The connections found by the system described in this paper are more modest. For instance, it has created formulas for testing prime numbers based both on Fermat's little theorem

Integer sequences are of central importance to the modeling of concepts admitting complete finitary descriptions. We introduce a novel view on the learning of such concepts and lay down a set of benchmarking tasks aimed at conceptual understanding by machine learning models. These tasks indirectly assess model ability to abstract, and challenge them to reason both interpolatively and extrapolatively from the knowledge gained by observing representative examples. To further aid research in knowledge representation and reasoning, we present FACT, the Finitary Abstraction Comprehension Toolkit.

We explore the possibility of using machine learning to identify interesting mathematical structures by using certain quantities that serve as fingerprints. In particular, we extract features from integer sequences using two empirical laws: Benford's law and Taylor's law and experiment with various classifiers to identify whether a sequence is nice, important, multiplicative, easy to compute or related to primes or palindromes.