Leveraging Online Olympiad-Level Math Problems for LLMs Training and Contamination-Resistant Evaluation
This work addresses the problem of unreliable evaluation and dataset scarcity for advanced math reasoning in LLMs, offering a scalable solution for researchers and developers, though it is incremental in automating data extraction from existing sources.
The authors tackled the limited size and quality of datasets for training and evaluating LLMs on Olympiad-level math problems by creating AoPS-Instruct, a dataset of over 600,000 QA pairs extracted from the Art of Problem Solving forum, and introduced LiveAoPSBench, a contamination-resistant benchmark that shows a significant decline in LLM performance over time, suggesting pre-training exposure rather than true reasoning ability.
Advances in Large Language Models (LLMs) have sparked interest in their ability to solve Olympiad-level math problems. However, the training and evaluation of these models are constrained by the limited size and quality of available datasets, as creating large-scale data for such advanced problems requires extensive effort from human experts. In addition, current benchmarks are prone to contamination, leading to unreliable evaluations. In this paper, we present an automated pipeline that leverages the rich resources of the Art of Problem Solving (AoPS) forum, which predominantly features Olympiad-level problems and community-driven solutions. Using open-source LLMs, we develop a method to extract question-answer pairs from the forum, resulting in AoPS-Instruct, a dataset of more than 600,000 high-quality QA pairs. Our experiments demonstrate that fine-tuning LLMs on AoPS-Instruct improves their reasoning abilities across various benchmarks. Moreover, we build an automatic pipeline that introduces LiveAoPSBench, an evolving evaluation set with timestamps, derived from the latest forum data, providing a contamination-resistant benchmark for assessing LLM performance. Notably, we observe a significant decline in LLM performance over time, suggesting their success on older examples may stem from pre-training exposure rather than true reasoning ability. Our work presents a scalable approach to creating and maintaining large-scale, high-quality datasets for advanced math reasoning, offering valuable insights into the capabilities and limitations of LLMs in this domain. Our benchmark and code is available at https://github.com/DSL-Lab/aops