Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Teaching-Inspired Integrated Prompting Framework: A Novel Approach for Enhancing Reasoning in Large Language Models

Wenting Tan, Dongxiao Chen, Jieting Xue, Zihao Wang, Taijie Chen

TL;DR

A novel and effective Teaching-Inspired Integrated Framework, which emulates the instructional process of a teacher guiding students, and equips LLMs with essential concepts, relevant theorems, and similar problems with analogous solution approaches, facilitating the enhancement of reasoning abilities.

Abstract

Large Language Models (LLMs) exhibit impressive performance across various domains but still struggle with arithmetic reasoning tasks. Recent work shows the effectiveness of prompt design methods in enhancing reasoning capabilities. However, these approaches overlook crucial requirements for prior knowledge of specific concepts, theorems, and tricks to tackle most arithmetic reasoning problems successfully. To address this issue, we propose a novel and effective Teaching-Inspired Integrated Framework, which emulates the instructional process of a teacher guiding students. This method equips LLMs with essential concepts, relevant theorems, and similar problems with analogous solution approaches, facilitating the enhancement of reasoning abilities. Additionally, we introduce two new Chinese datasets, MathMC and MathToF, both with detailed explanations and answers. Experiments are conducted on nine benchmarks which demonstrates that our approach improves the reasoning accuracy of LLMs. With GPT-4 and our framework, we achieve new state-of-the-art performance on four math benchmarks (AddSub, SVAMP, Math23K and AQuA) with accuracies of 98.2% (+3.3%), 93.9% (+0.2%), 94.3% (+7.2%) and 81.1% (+1.2%). Our data and code are available at https://github.com/SallyTan13/Teaching-Inspired-Prompting.

Teaching-Inspired Integrated Prompting Framework: A Novel Approach for Enhancing Reasoning in Large Language Models

TL;DR

A novel and effective Teaching-Inspired Integrated Framework, which emulates the instructional process of a teacher guiding students, and equips LLMs with essential concepts, relevant theorems, and similar problems with analogous solution approaches, facilitating the enhancement of reasoning abilities.

Abstract

Large Language Models (LLMs) exhibit impressive performance across various domains but still struggle with arithmetic reasoning tasks. Recent work shows the effectiveness of prompt design methods in enhancing reasoning capabilities. However, these approaches overlook crucial requirements for prior knowledge of specific concepts, theorems, and tricks to tackle most arithmetic reasoning problems successfully. To address this issue, we propose a novel and effective Teaching-Inspired Integrated Framework, which emulates the instructional process of a teacher guiding students. This method equips LLMs with essential concepts, relevant theorems, and similar problems with analogous solution approaches, facilitating the enhancement of reasoning abilities. Additionally, we introduce two new Chinese datasets, MathMC and MathToF, both with detailed explanations and answers. Experiments are conducted on nine benchmarks which demonstrates that our approach improves the reasoning accuracy of LLMs. With GPT-4 and our framework, we achieve new state-of-the-art performance on four math benchmarks (AddSub, SVAMP, Math23K and AQuA) with accuracies of 98.2% (+3.3%), 93.9% (+0.2%), 94.3% (+7.2%) and 81.1% (+1.2%). Our data and code are available at https://github.com/SallyTan13/Teaching-Inspired-Prompting.

Paper Structure

This paper contains 29 sections, 4 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. In-context Learning
  4. Reasoning with Prompting
  5. MathMC and MathToF
  6. Teaching-Inspired Integrated Prompting Framework
  7. Teaching-Inspired Prompts Generation
  8. Answers Generation
  9. Answers Selection
  10. Experiments
  11. Experimental Settings
  12. Main Results
  13. Ablation Study
  14. Similar problems and background knowledge.
  15. Python programs and double-check verification strategy.
  16. ...and 14 more sections

Figures (4)

  • Figure 1: Architecture of our Teaching-Inspired Integrated Prompting Framework
  • Figure 2: Experiment results of adding different number of varying or the same similar problems.
  • Figure 3: The procedure of similar problems retrieval (a) and the background knowledge generation (b).
  • Figure 4: Sample questions of MathMC and MathToF datasets. The left side displays the original questions in Chinese, while the right side shows the same questions translated into English