AS-ES Learning: Towards efficient CoT learning in small models

Nuwa Xi, Yuhan Chen, Sendong Zhao, Haochun Wang, GongZhang GongZhang, Bing Qin, Ting Liu


Abstract
Chain-of-Thought (CoT) serves as a critical emerging ability in LLMs, especially when it comes to logical reasoning. Attempts have been made to induce such ability in small models as well by distilling from the data with CoT generated by Large Language Models (LLMs). However, existing methods often simply generate and incorporate more data from LLMs and fail to note the importance of efficiently utilizing existing CoT data. We here propose a new training paradigm AS-ES (Abstractive Segments - Extractive Segments) learning, which exploits the inherent information in CoT for iterative generation. Experiments show that our methods surpass the direct seq2seq training on CoT-extensive tasks like MWP and PET summarization, without data augmentation or altering the model itself. Furthermore, we explore the reason behind the inefficiency of small models in learning CoT and provide an explanation of why AS-ES learning works, giving insights into the underlying mechanism of CoT.
Anthology ID:
2024.findings-acl.635
Volume:
Findings of the Association for Computational Linguistics: ACL 2024
Month:
August
Year:
2024
Address:
Bangkok, Thailand
Editors:
Lun-Wei Ku, Andre Martins, Vivek Srikumar
Venue:
Findings
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
10686–10697
Language:
URL:
https://preview.aclanthology.org/build-pipeline-with-new-library/2024.findings-acl.635/
DOI:
10.18653/v1/2024.findings-acl.635
Bibkey:
Cite (ACL):
Nuwa Xi, Yuhan Chen, Sendong Zhao, Haochun Wang, GongZhang GongZhang, Bing Qin, and Ting Liu. 2024. AS-ES Learning: Towards efficient CoT learning in small models. In Findings of the Association for Computational Linguistics: ACL 2024, pages 10686–10697, Bangkok, Thailand. Association for Computational Linguistics.
Cite (Informal):
AS-ES Learning: Towards efficient CoT learning in small models (Xi et al., Findings 2024)
Copy Citation:
PDF:
https://preview.aclanthology.org/build-pipeline-with-new-library/2024.findings-acl.635.pdf