https://github.com/IntelLabs/RAG-FiT/tree/square.

\n","updatedAt":"2025-02-14T07:35:53.727Z","author":{"_id":"62d93cd728f9c86a4031562e","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/62d93cd728f9c86a4031562e/ix_LD-wjW8vCltosCVUmV.jpeg","fullname":"Daniel Fleischer","name":"danf","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":11,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.8791781663894653},"editors":["danf"],"editorAvatarUrls":["https://cdn-avatars.huggingface.co/v1/production/uploads/62d93cd728f9c86a4031562e/ix_LD-wjW8vCltosCVUmV.jpeg"],"reactions":[],"isReport":false}},{"id":"67afeefb404cc27c6d9f8349","author":{"_id":"63d3e0e8ff1384ce6c5dd17d","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1674830754237-63d3e0e8ff1384ce6c5dd17d.jpeg","fullname":"Librarian Bot (Bot)","name":"librarian-bot","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":318,"isUserFollowing":false},"createdAt":"2025-02-15T01:33:47.000Z","type":"comment","data":{"edited":false,"hidden":false,"latest":{"raw":"This is an automated message from the [Librarian Bot](https://huggingface.co/librarian-bots). I found the following papers similar to this paper. \n\nThe following papers were recommended by the Semantic Scholar API \n\n* [Bactrainus: Optimizing Large Language Models for Multi-hop Complex Question Answering Tasks](https://huggingface.co/papers/2501.06286) (2025)\n* [ARR: Question Answering with Large Language Models via Analyzing, Retrieving, and Reasoning](https://huggingface.co/papers/2502.04689) (2025)\n* [Understanding Before Reasoning: Enhancing Chain-of-Thought with Iterative Summarization Pre-Prompting](https://huggingface.co/papers/2501.04341) (2025)\n* [IAO Prompting: Making Knowledge Flow Explicit in LLMs through Structured Reasoning Templates](https://huggingface.co/papers/2502.03080) (2025)\n* [Recursive Decomposition of Logical Thoughts: Framework for Superior Reasoning and Knowledge Propagation in Large Language Models](https://huggingface.co/papers/2501.02026) (2025)\n* [Prompting Large Language Models with Rationale Heuristics for Knowledge-based Visual Question Answering](https://huggingface.co/papers/2412.16936) (2024)\n* [Knowledge Editing through Chain-of-Thought](https://huggingface.co/papers/2412.17727) (2024)\n\n\n Please give a thumbs up to this comment if you found it helpful!\n\n If you want recommendations for any Paper on Hugging Face checkout [this](https://huggingface.co/spaces/librarian-bots/recommend_similar_papers) Space\n\n You can directly ask Librarian Bot for paper recommendations by tagging it in a comment: `@librarian-bot recommend`","html":"

This is an automated message from the Librarian Bot. I found the following papers similar to this paper.

The following papers were recommended by the Semantic Scholar API

Please give a thumbs up to this comment if you found it helpful!

If you want recommendations for any Paper on Hugging Face checkout this Space

You can directly ask Librarian Bot for paper recommendations by tagging it in a comment: \n\n@librarian-bot\n\t recommend

\n","updatedAt":"2025-02-15T01:33:47.518Z","author":{"_id":"63d3e0e8ff1384ce6c5dd17d","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1674830754237-63d3e0e8ff1384ce6c5dd17d.jpeg","fullname":"Librarian Bot (Bot)","name":"librarian-bot","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":318,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.7404724359512329},"editors":["librarian-bot"],"editorAvatarUrls":["https://cdn-avatars.huggingface.co/v1/production/uploads/1674830754237-63d3e0e8ff1384ce6c5dd17d.jpeg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2502.09390","authors":[{"_id":"67aef17da9f929ce0ca3e36b","user":{"_id":"62d93cd728f9c86a4031562e","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/62d93cd728f9c86a4031562e/ix_LD-wjW8vCltosCVUmV.jpeg","isPro":false,"fullname":"Daniel Fleischer","user":"danf","type":"user"},"name":"Daniel Fleischer","status":"extracted_pending","statusLastChangedAt":"2025-02-14T07:32:14.019Z","hidden":false},{"_id":"67aef17da9f929ce0ca3e36c","user":{"_id":"63e0c8875c6964861ebb0c49","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/63e0c8875c6964861ebb0c49/yzkhPSxgXtJCM62iMBOOK.jpeg","isPro":false,"fullname":"Moshe Berchansky","user":"mber","type":"user"},"name":"Moshe Berchansky","status":"admin_assigned","statusLastChangedAt":"2025-02-14T12:52:07.600Z","hidden":false},{"_id":"67aef17da9f929ce0ca3e36d","user":{"_id":"666154a2694ec45eaabc19c3","avatarUrl":"/avatars/b0022ad6dd992c75a51973070b302315.svg","isPro":false,"fullname":"Gad Markovits","user":"gadmarkovits","type":"user"},"name":"Gad Markovits","status":"admin_assigned","statusLastChangedAt":"2025-02-14T12:52:13.549Z","hidden":false},{"_id":"67aef17da9f929ce0ca3e36e","user":{"_id":"6013c64031d84a3bb7650038","avatarUrl":"/avatars/e05ca715004b39e79472399f75010bda.svg","isPro":false,"fullname":"Moshe Wasserblat","user":"moshew","type":"user"},"name":"Moshe Wasserblat","status":"admin_assigned","statusLastChangedAt":"2025-02-14T12:52:19.452Z","hidden":false}],"publishedAt":"2025-02-13T15:07:20.000Z","submittedOnDailyAt":"2025-02-14T05:05:53.718Z","title":"SQuARE: Sequential Question Answering Reasoning Engine for Enhanced\n Chain-of-Thought in Large Language Models","submittedOnDailyBy":{"_id":"62d93cd728f9c86a4031562e","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/62d93cd728f9c86a4031562e/ix_LD-wjW8vCltosCVUmV.jpeg","isPro":false,"fullname":"Daniel Fleischer","user":"danf","type":"user"},"summary":"In the rapidly evolving field of Natural Language Processing, Large Language\nModels (LLMs) are tasked with increasingly complex reasoning challenges.\nTraditional methods like chain-of-thought prompting have shown promise but\noften fall short in fully leveraging a model's reasoning capabilities. This\npaper introduces SQuARE (Sequential Question Answering Reasoning Engine), a\nnovel prompting technique designed to improve reasoning through a\nself-interrogation paradigm. Building upon CoT frameworks, SQuARE prompts\nmodels to generate and resolve multiple auxiliary questions before tackling the\nmain query, promoting a more thorough exploration of various aspects of a\ntopic. Our expansive evaluations, conducted with Llama 3 and GPT-4o models\nacross multiple question-answering datasets, demonstrate that SQuARE\nsignificantly surpasses traditional CoT prompts and existing\nrephrase-and-respond methods. By systematically decomposing queries, SQuARE\nadvances LLM capabilities in reasoning tasks. The code is publicly available at\nhttps://github.com/IntelLabs/RAG-FiT/tree/square.","upvotes":16,"discussionId":"67aef17ea9f929ce0ca3e3bf","githubRepo":"https://github.com/IntelLabs/RAG-FiT","githubRepoAddedBy":"auto","ai_summary":"SQuARE, a novel prompting technique, enhances Large Language Models' reasoning by generating and resolving multiple auxiliary questions before addressing the main query, achieving better results than traditional methods.","ai_keywords":["Large Language Models (LLMs)","chain-of-thought prompting","Sequential Question Answering Reasoning Engine (SQuARE)","CoT frameworks","Llama 3","GPT-4o","question-answering datasets","rephrase-and-respond methods"],"githubStars":767},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"62d93cd728f9c86a4031562e","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/62d93cd728f9c86a4031562e/ix_LD-wjW8vCltosCVUmV.jpeg","isPro":false,"fullname":"Daniel Fleischer","user":"danf","type":"user"},{"_id":"6055ae5d25cd24537dd59dc5","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/6055ae5d25cd24537dd59dc5/eswozkCirLrnyhufN8_-f.jpeg","isPro":false,"fullname":"Daniel Korat","user":"danielkorat","type":"user"},{"_id":"6013c64031d84a3bb7650038","avatarUrl":"/avatars/e05ca715004b39e79472399f75010bda.svg","isPro":false,"fullname":"Moshe Wasserblat","user":"moshew","type":"user"},{"_id":"648eb1eb59c4e5c87dc116e0","avatarUrl":"/avatars/c636cea39c2c0937f01398c94ead5dad.svg","isPro":false,"fullname":"fdsqefsgergd","user":"T-representer","type":"user"},{"_id":"650c8bfb3d3542884da1a845","avatarUrl":"/avatars/863a5deebf2ac6d4faedc4dd368e0561.svg","isPro":false,"fullname":"Adhurim ","user":"Limi07","type":"user"},{"_id":"66f612b934b8ac9ffa44f084","avatarUrl":"/avatars/6836c122e19c66c90f1673f28b30d7f0.svg","isPro":false,"fullname":"Tang","user":"tommysally","type":"user"},{"_id":"63e0c8875c6964861ebb0c49","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/63e0c8875c6964861ebb0c49/yzkhPSxgXtJCM62iMBOOK.jpeg","isPro":false,"fullname":"Moshe Berchansky","user":"mber","type":"user"},{"_id":"67b0a4aa8191c180b9421ba7","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/67b0a4aa8191c180b9421ba7/xg3KQqy5nlqbaoIjxxRhJ.png","isPro":false,"fullname":"Taro","user":"KDoyoon","type":"user"},{"_id":"62f32eab52ad88c930bb3f3b","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1677134945205-62f32eab52ad88c930bb3f3b.png","isPro":false,"fullname":"Asankhaya Sharma","user":"codelion","type":"user"},{"_id":"65afe3fd7c11edbf6e1a1277","avatarUrl":"/avatars/a35c3e29d712d1fb062b5eb8887d46a6.svg","isPro":false,"fullname":"Robin Williams","user":"bfuzzy1","type":"user"},{"_id":"606d6349f1259f30578520ad","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/606d6349f1259f30578520ad/72_XrFfgQ6p9tgJj5Bc5U.png","isPro":false,"fullname":"Jonathan Mamou","user":"jmamou","type":"user"},{"_id":"651c80a26ba9ab9b9582c273","avatarUrl":"/avatars/e963452eafd21f517d800f2e58e0f918.svg","isPro":false,"fullname":"siyeng feng","user":"siyengfeng","type":"user"}],"acceptLanguages":["*"],"dailyPaperRank":0}">

Papers

arxiv:2502.09390

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

Published on Feb 13, 2025

· Submitted by

Daniel Fleischer on Feb 14, 2025

Upvote

Authors:

Daniel Fleischer ,

Moshe Berchansky ,

Gad Markovits ,

Moshe Wasserblat

Abstract

SQuARE, a novel prompting technique, enhances Large Language Models' reasoning by generating and resolving multiple auxiliary questions before addressing the main query, achieving better results than traditional methods.

AI-generated summary

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 arXiv page View PDF GitHub 767 auto Add to collection

Community

danf

Paper author Paper submitter Feb 14, 2025

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.