@inproceedings{8edd001fdfb94f41b6c9c4152bf0f14c,
title = "Masked Thought: Simply Masking Partial Reasoning Steps Can Improve Mathematical Reasoning Learning of Language Models",
abstract = "In reasoning tasks, even a minor error can cascade into inaccurate results, leading to suboptimal performance of large language models in such domains. Earlier fine-tuning approaches sought to mitigate this by leveraging more precise supervisory signals from human labeling, larger models, or self-sampling, although at a high cost. Conversely, we develop a method that avoids external resources, relying instead on introducing perturbations to the input. Our training approach randomly masks certain tokens within the chain of thought, a technique we found to be particularly effective for reasoning tasks. When applied to fine-tuning with GSM8K on Llama-2-7B, this method achieved a 5% improvement in GSM8K accuracy and a 10% improvement in GSM-IC accuracy over standard supervised fine-tuning with a few codes modified. Furthermore, it is complementary to existing methods. When integrated with related explicit data augmentation methods, it leads to improvements across five datasets of various augmentation methods, as well as two different base models. We further investigate the mechanisms behind this improvement through case studies and quantitative analysis, suggesting that our approach may provide superior support for the model in capturing long-distance dependencies, especially those related to questions. This enhancement could deepen understanding of the premises in questions and prior steps. Our code is available at Github.",
author = "Changyu Chen and Xiting Wang and Lin, {Ting En} and Ang Lv and Yuchuan Wu and Xin Gao and Wen, {Ji Rong} and Rui Yan and Yongbin Li",
note = "Publisher Copyright: {\textcopyright} 2024 Association for Computational Linguistics.; 62nd Annual Meeting of the Association for Computational Linguistics, ACL 2024 ; Conference date: 11-08-2024 Through 16-08-2024",
year = "2024",
language = "English (US)",
series = "Proceedings of the Annual Meeting of the Association for Computational Linguistics",
publisher = "Association for Computational Linguistics (ACL)",
pages = "5872--5900",
editor = "Lun-Wei Ku and Martins, {Andre F. T.} and Vivek Srikumar",
booktitle = "Long Papers",
address = "United States",
}