SQuARE: Sequential Question Answering Reasoning Engine for Enhanced Chain-of-Thought in Large Language Models

Daniel Fleischer, Moshe Berchansky, Gad Markovits, Moshe Wasserblat
2/13/2025
cs.CLcs.AIcs.LG

Official code: IntelLabs/RAG-FiT (771 stars, Apache-2.0, last commit 2025-12-16).

Abstract

In the rapidly evolving field of Natural Language Processing, Large Language Models (LLMs) are tasked with increasingly complex reasoning challenges. Traditional methods like chain-of-thought prompting have shown promise but often fall short in fully leveraging a model's reasoning capabilities. This paper introduces SQuARE (Sequential Question Answering Reasoning Engine), a novel prompting technique designed to improve reasoning through a self-interrogation paradigm. Building upon CoT frameworks, SQuARE prompts models to generate and resolve multiple auxiliary questions before tackling the main query, promoting a more thorough exploration of various aspects of a topic. Our expansive evaluations, conducted with Llama 3 and GPT-4o models across multiple question-answering datasets, demonstrate that SQuARE significantly surpasses traditional CoT prompts and existing rephrase-and-respond methods. By systematically decomposing queries, SQuARE advances LLM capabilities in reasoning tasks. The code is publicly available at https://github.com/IntelLabs/RAG-FiT/tree/square.

View on arXivView PDF

Code Implementations(1)

IntelLabs/RAG-FiTOfficial100%
Apache-2.0

Framework for enhancing LLMs for RAG tasks using fine-tuning.

77161PythonJul 23, 202410 months agoApache-2.0
evaluationfine-tuninginformation-retrievalllmnlp+3 more

Similar Papers

Graph Elicitation for Guiding Multi-Step Reasoning in Large Language Models

83% match2023

Seek and Solve Reasoning for Table Question Answering

83% match2024

Question-Analysis Prompting Improves LLM Performance in Reasoning Tasks

83% match2024

Interactive-KBQA: Multi-Turn Interactions for Knowledge Base Question Answering with Large Language Models

83% match1 repo302024

SG-FSM: A Self-Guiding Zero-Shot Prompting Paradigm for Multi-Hop Question Answering Based on Finite State Machine

83% match2024

Iteration of Thought: Leveraging Inner Dialogue for Autonomous Large Language Model Reasoning

83% match1 repo1352024

Cite this paper

@article{fleischer2025square,
  title  = {SQuARE: Sequential Question Answering Reasoning Engine for Enhanced Chain-of-Thought in Large Language Models},
  author = {Daniel Fleischer and Moshe Berchansky and Gad Markovits and Moshe Wasserblat},
  year   = {2025},
  eprint = {2502.09390},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2502.09390v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2502.09390.svg)](https://codeofpaper.com/paper/2502.09390)

Discussion