DeFT-RAG:將多跳問題拆解為可重用答案的自包含子問題
DeFT-RAG: Decomposing Multi-Hop Questions into Self-Contained Sub-Questions with Reusable Answers
DeFT-RAG 透過逐步生成自包含子問題、以規則更新和輕量決策模型處理部分工作,以及快取和重用答案,改善多跳問答。在 MuSiQue、2WikiMultiHopQA 和 HotpotQA 上,使用 gpt-4o-mini 和 Qwen3.6 作為基礎模型時,其準確率均勝過 13 個基準方法,輸入 LLM 的 token 數亦少於其他迭代式方法。
Published on Oct 30
Authors:
Abstract
DeFT-RAG improves multi-hop question answering by generating self-contained sub-questions sequentially, offloading reasoning to rule-based updates and a lightweight decision model, and caching answers to reduce LLM usage.
Retrieval-augmented generation (RAG) lets a large language model (LLM) answer questions by retrieving relevant passages from a large corpus. However, RAG struggles with multi-hop questions, which require chaining facts spread across several documents through intermediate entities that the question never mentions. Graph-based RAG methods retrieve once from the entities a question names and so rarely recover the full evidence chain. Iterative RAG methods recover it step by step, but their accuracy is limited by the many tasks they ask the LLM to perform, such as planning the sub-questions, deciding their dependencies, rewriting them as answers arrive, and keeping track of what is known. We propose DeFT-RAG, a RAG framework built on three components. First, DeFT-RAG generates one self-contained sub-question at a time from the original question and the sub-questions answered so far with their answers (QA pairs). Each sub-question names every entity it asks about, which removes the need to plan, track dependencies or rewrite, and lets the LLM focus on a single fine-grained task at a time. Second, DeFT-RAG moves the remaining LLM work outside the LLM: rule-based updates to a QA table that keeps track of what is known, and yes/no decisions made by a fine-tuned lightweight decision model. Third, DeFT-RAG caches the QA pairs and reuses an answer when a later question asks a sub-question with the same meaning. On MuSiQue, 2WikiMultiHopQA and HotpotQA, DeFT-RAG achieves the best accuracy among thirteen baselines with both gpt-4o-mini and Qwen3.6 as the backbone, while giving the LLM fewer input tokens than any other iterative method. With a QA cache shared across 1,000 MuSiQue questions, 23.1% of sub-questions are answered from the cache, which reduces LLM calls by 12.9% without lowering accuracy.
View arXiv page View PDF GitHub 1 Add to collection
Get this paper in your agent:
hf papers read 2610.37826
Don't have the latest CLI?
curl -LsSf https://hf.co/cli/install.sh | bash
Models citing this paper 1
5B • Updated about 6 hours ago • 1Dingy-Kang/DeFT-RAG-DM-4B
Datasets citing this paper 0
No dataset linking this paper
Cite arxiv.org/abs/2610.37826 in a dataset README.md to link it from this page.
Spaces citing this paper 0
No Space linking this paper
Cite arxiv.org/abs/2610.37826 in a Space README.md to link it from this page.
Collections including this paper 0
No Collection including this paper
Add this paper to a collection to link it from this page.
來源:HuggingFace Daily Papers(社區熱門論文) · huggingface.co