mlsanityData to 14 Oct 2026

Only Ask What You Don't Know: Grounded Delta Planning for Efficient Multi-step RAG

arXiv 2606.22681 ↗First posted 21 Jun 2026cs.CL cs.AIcited by 3

On the map: LLM reasoning → Knowledge- and tool-augmented reasoning

also in 1 more
Abstract (arXiv)

Multi-hop question answering remains challenging for Retrieval-Augmented Generation (RAG) because existing approaches either propagate errors across iterative retrieval rounds or over-generate reasoning steps, increasing cost without improving accuracy. We propose Grounded Delta Planning RAG (GDP-RAG), a plan-based framework that targets only the information delta based on three simple design choices: (1) preliminary retrieval to ground planning before execution, (2) a gap-conditioned planning prompt that asks only for missing information, and (3) a skeletal trajectory that pairs each subquery with a Thought capturing evidence from preliminary retrieval and carrying it through to the final answer. GDP-RAG focuses computation on unresolved gaps, yielding concise, reliable reasoning trajectories. Extensive experiments on HotpotQA, 2WikiMultiHopQA, and MuSiQue show that GDP-RAG achieves the highest accuracy (60.63%) among all compared systems while maintaining a cost-of-pass of 0.51, 22% lower than PAR-RAG (0.65) and 68% lower than KnowTrace (1.57), with no method achieving both higher accuracy and lower cost.

Key facts. Cited by 3 papers in the map (data through 14 Oct 2026). Placed in Knowledge- and tool-augmented reasoning, one of 17 directions in LLM reasoning. Builds on 44 of its 47 references that are in the map. First posted on arXiv 21 Jun 2026.

Builds on

44 of 47 references.

Closest ancestors · 8

Same direction, newest first

  1. 2025MA-RAG: Multi-Agent Retrieval-Augmented Generation via Collaborative Chain-of-Thought Reasoning26 May
  2. 2025Search-R1: Training LLMs to Reason and Leverage Search Engines with Reinforcement Learning12 Mar
  3. 2025R1-Searcher: Incentivizing the Search Capability in LLMs via Reinforcement Learning7 Mar
  4. 2025Search-o1: Agentic Search-Enhanced Large Reasoning Models9 Jan
  5. 2024RAT: Retrieval Augmented Thoughts Elicit Context-Aware Reasoning in Long-Horizon Generation8 Mar
  6. 2023Retrieval-Augmented Generation for Large Language Models: A Survey18 Dec
  7. 2023Leveraging Structured Information for Explainable Multi-hop Question Answering and Reasoning7 Nov
  8. 2022Interleaving Retrieval with Chain-of-Thought Reasoning for Knowledge-Intensive Multi-Step Questions20 Dec

Other references · 36

  1. 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Modelscited by 12,459
  2. 2020Retrieval-Augmented Generation for Knowledge-Intensive NLP Taskscited by 9,297
  3. 2022ReAct: Synergizing Reasoning and Acting in Language Modelscited by 8,695
  4. 2022Large Language Models are Zero-Shot Reasonerscited by 4,374
  5. 2018HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answeringcited by 3,754
  6. 2023Lost in the Middle: How Language Models Use Long Contextscited by 3,455
  7. 2023Self-RAG: Learning to Retrieve, Generate, and Critique through Self-Reflectioncited by 1,839
  8. 2024From Local to Global: A Graph RAG Approach to Query-Focused Summarizationcited by 1,453
  9. 2020Constructing A Multi-hop QA Dataset for Comprehensive Evaluation of Reasoning Stepscited by 1,282
  10. 2022Measuring and Narrowing the Compositionality Gap in Language Modelscited by 1,075
  11. 2024M3-Embedding: Multi-Linguality, Multi-Functionality, Multi-Granularity Text Embeddings Through Self-Knowledge Distillationcited by 1,073
  12. 2021MuSiQue: Multihop Questions via Single-hop Question Compositioncited by 916
24 more
  1. 2022Precise Zero-Shot Dense Retrieval without Relevance Labelscited by 517
  2. 2023Faith and Fate: Limits of Transformers on Compositionalitycited by 503
  3. 2018The Web as a Knowledge-base for Answering Complex Questionscited by 410
  4. 2023Enhancing Retrieval-Augmented Large Language Models with Iterative Retrieval-Generation Synergycited by 346
  5. 2023Chain-of-Verification Reduces Hallucination in Large Language Modelscited by 324
  6. 2024HippoRAG: Neurobiologically Inspired Long-Term Memory for Large Language Modelscited by 281
  7. 2025Agentic Retrieval-Augmented Generation: A Survey on Agentic RAGcited by 249
  8. 2024Corrective Retrieval Augmented Generationcited by 197
  9. 2017Constructing Datasets for Multi-hop Reading Comprehension Across Documentscited by 190
  10. 2023ReWOO: Decoupling Reasoning from Observations for Efficient Augmented Language Modelscited by 184
  11. 2024Assisting in Writing Wikipedia-like Articles From Scratch with Large Language Modelscited by 154
  12. 2025Efficient Agents: Building Effective Agents While Reducing Costcited by 32
  13. 2024Knowledge-Aware Query Expansion with Large Language Models for Textual and Relational Retrievalcited by 26
  14. 2024Citation-Enhanced Generation for LLM-based Chatbotscited by 25
  15. 2024Learning to Plan for Retrieval-Augmented Large Language Models from Knowledge Graphscited by 24
  16. 2024ERA-CoT: Improving Chain-of-Thought through Entity Relationship Analysiscited by 18
  17. 2025RAS: Retrieval-And-Structuring for Knowledge-Intensive LLM Generationcited by 16
  18. 2025Credible Plan-Driven RAG Method for Multi-Hop Question Answeringcited by 12
  19. 2025ComposeRAG: A Modular and Composable RAG for Corpus-Grounded Multi-Hop Question Answeringcited by 11
  20. 2025KnowTrace: Bootstrapping Iterative Retrieval-Augmented Generation with Structured Knowledge Tracingcited by 8
  21. 2024Analysis of Plan-based Retrieval for Grounded Text Generationcited by 5
  22. 2025A Preliminary Study of RAG for Taiwanese Historical Archivescited by 4
  23. 2026MultiCube-RAG for Multi-hop Question Answeringcited by 3
  24. 2026CodaRAG: Connecting the Dots with Associativity Inspired by Complementary Learningcited by 3

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 3.

3 papers.

  1. 2026Bridging Static and Agentic RAG for Taiwanese Historical Question Answeringcited by 0
  2. 2026MetaBench-Harness: Unlocking End-to-End Optimization of Benchmark Harnessescited by 0
  3. 2026TAEC: Trajectory-Aware Evidence Coordination for Multi-Step Visual RAGcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025Investigating Retrieval-Augmented Generation Systems on Unanswerable, Uncheatable, Realistic, Multi-hop Queries
  2. 2025REAP: Enhancing RAG with Recursive Evaluation and Adaptive Planning for Multi-Hop Question Answering
  3. 2025Think Straight, Stop Smart: Structured Reasoning for Efficient Multi-Hop RAG
  4. 2024Plan*RAG: Efficient Test-Time Planning for Retrieval Augmented Generation
  5. 2026PAR2-RAG: Planned Active Retrieval and Reasoning for Multi-Hop Question Answering
  6. 2026Reasoning in Trees: Improving Retrieval-Augmented Generation for Multi-Hop Question Answering
  7. 2025Memory-Aware and Uncertainty-Guided Retrieval for Multi-Hop Question Answering
  8. 2026CompactRAG: Reducing LLM Calls and Token Overhead in Multi-Hop Question Answering