Recent advances in large language models (LLMs) have demonstrated the effectiveness of Iterative Self-Improvement (ISI) techniques. However, continuous training on self-generated data leads to reduced output diversity, a limitation particularly critical in reasoning tasks where diverse solution paths are essential. We present DIVE (Diversified Iterative Self-Improvement), a novel framework that addresses this challenge through two key components: Sample Pool Expansion for broader solution exploration, and Data Selection for balancing diversity and quality in preference pairs. Experiments on MATH and GSM8k datasets show that DIVE achieves a 10% to 45% relative increase in output diversity metrics while maintaining performance quality compared to vanilla ISI. Our ablation studies confirm both components' significance in achieving these improvements. Code is available at https://github.com/qinyiwei/DIVE.
@article{arxiv.2501.00747,
title = {DIVE: Diversified Iterative Self-Improvement},
author = {Yiwei Qin and Yixiu Liu and Pengfei Liu},
journal= {arXiv preprint arXiv:2501.00747},
year = {2025}
}