AS-ES Learning: Towards Efficient CoT Learning in Small Models
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2024
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866910353192910848 |
|---|---|
| author | Xi, Nuwa Chen, Yuhan Zhao, Sendong Wang, Haochun Qin, Bing Liu, Ting |
| author_facet | Xi, Nuwa Chen, Yuhan Zhao, Sendong Wang, Haochun Qin, Bing Liu, Ting |
| contents | 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. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2403_01969 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | AS-ES Learning: Towards Efficient CoT Learning in Small Models Xi, Nuwa Chen, Yuhan Zhao, Sendong Wang, Haochun Qin, Bing Liu, Ting Computation and Language 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. |
| title | AS-ES Learning: Towards Efficient CoT Learning in Small Models |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2403.01969 |