mlsanityData to 14 Oct 2026

CoVeR: Coverage-Based Routing of Verifier Calls in Agentic Retrieval

arXiv 2609.26086 ↗First posted 3 Sep 2026cs.CL cs.IRcited by 1

On the map: LLM agents and tool use → Web, search, and research agents

Abstract (arXiv)

An agentic retrieval system issues a sequence of search queries and must decide, at each step, whether the evidence collected so far is enough to stop. Delegating that decision to an LLM verifier or a prompt judge makes stopping reliable, but the verifier then reprocesses the growing evidence after every retrieval step, a substantial repeated cost. We show that most of these calls can be skipped without materially changing answer accuracy: a single threshold on a frozen sentence-embedding coverage margin detects the states in which the evidence is still plainly incomplete, and the verifier is called only on the ambiguous remainder, a gate we call CoVeR (Coverage-based Verifier Routing). Across three multi-hop QA benchmarks, with the evaluation protocol fixed before the full-scale run, the CoVeR-gated agent matches the answer accuracy of both the full-budget agent and the always-verify baseline within a fraction of an EM point. It cuts 62-68% of verifier calls, and 93% in a saturated regime. Routers built on evidence counts, lexical overlap, or BM25 relevance, alone or learned in combination, give weaker overall trade-offs, the gate transfers without re-tuning across deciders and agent scales, and its drafter distills into a 921k-parameter head atop the frozen encoder, leaving no LLM in the routing loop. The same signal cannot replace verification: matching a claim is far easier than deciding the claim is supported.

Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Placed in Web, search, and research agents, one of 16 directions in LLM agents and tool use. Builds on 30 of its 36 references that are in the map. First posted on arXiv 3 Sep 2026.

Builds on

30 of 36 references.

References · 30

  1. 2020Retrieval-Augmented Generation for Knowledge-Intensive NLP Taskscited by 9,297
  2. 2018HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answeringcited by 3,754
  3. 2023Self-RAG: Learning to Retrieve, Generate, and Critique through Self-Reflectioncited by 1,839
  4. 2022Fast Inference from Transformers via Speculative Decodingcited by 1,653
  5. 2025Search-R1: Training LLMs to Reason and Leverage Search Engines with Reinforcement Learningcited by 1,444
  6. 2018FEVER: a large-scale dataset for Fact Extraction and VERificationcited by 1,375
  7. 2020Constructing A Multi-hop QA Dataset for Comprehensive Evaluation of Reasoning Stepscited by 1,282
  8. 2022Text Embeddings by Weakly-Supervised Contrastive Pre-trainingcited by 1,253
  9. 2022Measuring and Narrowing the Compositionality Gap in Language Modelscited by 1,075
  10. 2023Accelerating Large Language Model Decoding with Speculative Samplingcited by 1,006
  11. 2022Interleaving Retrieval with Chain-of-Thought Reasoning for Knowledge-Intensive Multi-Step Questionscited by 923
  12. 2021MuSiQue: Multihop Questions via Single-hop Question Compositioncited by 916
18 more
  1. 2023FrugalGPT: How to Use Large Language Models While Reducing Cost and Improving Performancecited by 853
  2. 2023Enabling Large Language Models to Generate Text with Citationscited by 595
  3. 2023C-Pack: Packed Resources For General Chinese Embeddingscited by 578
  4. 2023Active Retrieval Augmented Generationcited by 516
  5. 2024Adaptive-RAG: Learning to Adapt Retrieval-Augmented Large Language Models through Question Complexitycited by 451
  6. 2020Consistent Estimators for Learning to Defer to an Expertcited by 271
  7. 2017Predict Responsibly: Improving Fairness and Accuracy by Learning to Defercited by 233
  8. 2025ReSearch: Learning to Reason with Search for LLMs via Reinforcement Learningcited by 161
  9. 2022Attributed Question Answering: Evaluation and Modeling for Attributed Large Language Modelscited by 112
  10. 2024Speculative RAG: Enhancing Retrieval Augmented Generation through Draftingcited by 77
  11. 2024Sufficient Context: A New Lens on Retrieval Augmented Generation Systemscited by 72
  12. 2025Knowing You Don't Know: Learning When to Continue Search in Multi-round RAG through Self-Practicingcited by 25
  13. 2025Stop-RAG: Value-Based Retrieval Control for Iterative RAGcited by 14
  14. 2026Adaptive Test-Time Compute Allocation via Learned Heuristics over Categorical Structurecited by 5
  15. 2025FrugalRAG: Less is More in RL Finetuning for Multi-Hop Question Answeringcited by 4
  16. 2026SAGE: A Novelty Gate for Efficient Memory Evolution in Agentic LLMscited by 3
  17. 2026HALT: Verification-Aware Stopping for Retrieval-Augmented Search Agentscited by 3
  18. 2026RASER: Recoverability-Aware Selective Escalation Router for Multi-Hop Question Answeringcited by 2

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 1.

1 paper.

  1. 2026What Can Component-Replacement Evidence Establish? A Critical Scoping Review of Local Decisions in LLM Agentscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025PRISM: Agentic Retrieval with LLMs for Multi-Hop Question Answering
  2. 2026Agentic-R: Learning to Retrieve for Agentic Search
  3. 2026SearchAtlas: Analyzing Agentic Search Strategies via Evidential Query Graphs
  4. 2026Evidence-Inference Reconstruction: When The Evidence Is Recalled But The Reasoning Goes Wrong
  5. 2026AgenticRAGTracer: A Hop-Aware Benchmark for Diagnosing Multi-Step Retrieval Reasoning in Agentic RAG
  6. 2026Agentic Systems as Boosting Weak Reasoning Models
  7. 2026FineVerify: Scaling Test-Time Compute with Fine-Grained Self-Verification for Agentic Search
  8. 2026Dissecting Agentic RAG: A Component Ablation for Multi-Hop QA with a Local 7B Model