# AS-ES Learning: Towards Efficient CoT Learning in Small Models

arXiv: https://arxiv.org/abs/2403.01969 · HTML: https://mlsanity.com/paper/2403.01969/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Prompting and chain-of-thought reasoning, one of 17 directions in LLM reasoning.
- Builds on 21 of its 30 references that are in the map.
- First posted on arXiv 4 Mar 2024.

## Placed in

- LLM reasoning / [Prompting and chain-of-thought reasoning](https://mlsanity.com/lr/prompting-and-chain-of-thought-reasoning/)

## Builds on (most cited first, 21 of 21)

- [PyTorch: An Imperative Style, High-Performance Deep Learning Library](https://mlsanity.com/paper/1912.01703/) (2019)
- [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://mlsanity.com/paper/1910.10683/) (2019)
- [Self-Consistency Improves Chain of Thought Reasoning in Language Models](https://mlsanity.com/paper/2203.11171/) (2022)
- [Scaling Instruction-Finetuned Language Models](https://mlsanity.com/paper/2210.11416/) (2022)
- [mT5: A massively multilingual pre-trained text-to-text transformer](https://mlsanity.com/paper/2010.11934/) (2020)
- [Can LLMs Express Their Uncertainty? An Empirical Evaluation of Confidence Elicitation in LLMs](https://mlsanity.com/paper/2306.13063/) (2023)
- [Multimodal Chain-of-Thought Reasoning in Language Models](https://mlsanity.com/paper/2302.00923/) (2023)
- [Distilling Step-by-Step! Outperforming Larger Language Models with Less Training Data and Smaller Model Sizes](https://mlsanity.com/paper/2305.02301/) (2023)
- [Large Language Models Are Reasoning Teachers](https://mlsanity.com/paper/2212.10071/) (2022)
- [Specializing Smaller Language Models towards Multi-Step Reasoning](https://mlsanity.com/paper/2301.12726/) (2023)
- [Distilling Reasoning Capabilities into Smaller Language Models](https://mlsanity.com/paper/2212.00193/) (2022)
- [Scalable Multi-Hop Relational Reasoning for Knowledge-Aware Question Answering](https://mlsanity.com/paper/2005.00646/) (2020)
- [RNNLogic: Learning Logic Rules for Reasoning on Knowledge Graphs](https://mlsanity.com/paper/2010.04029/) (2020)
- [Improving Coherence and Consistency in Neural Sequence Models with Dual-System, Neuro-Symbolic Reasoning](https://mlsanity.com/paper/2107.02794/) (2021)
- [Multi-hop Question Answering](https://mlsanity.com/paper/2204.09140/) (2022)
- [Learning Reasoning Strategies in End-to-End Differentiable Proving](https://mlsanity.com/paper/2007.06477/) (2020)
- [Self-Assembling Modular Networks for Interpretable Multi-Hop Reasoning](https://mlsanity.com/paper/1909.05803/) (2019)
- [MCC-KD: Multi-CoT Consistent Knowledge Distillation](https://mlsanity.com/paper/2310.14747/) (2023)
- [Sci-CoT: Leveraging Large Language Models for Enhanced Knowledge Distillation in Small Models for Scientific QA](https://mlsanity.com/paper/2308.04679/) (2023)
- [Multi-hop Inference for Question-driven Summarization](https://mlsanity.com/paper/2010.03738/) (2020)
- [Disentangling Extraction and Reasoning in Multi-hop Spatial Reasoning](https://mlsanity.com/paper/2310.16731/) (2023)

## Built on by (most cited first, 1 of 1)

- [Small Language Models: Survey, Measurements, and Insights](https://mlsanity.com/paper/2409.15790/) (2024)

## Abstract (arXiv)

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.