[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-85920-en":3,"doc-seo-85920-105":30,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":21,"is_downloadable":21,"audit_status":21,"page_count":22,"language":23,"language_code":24,"site_id":25,"html_lang":24,"table_of_contents":26,"faqs":27,"seo_title":13,"seo_description":14,"update_tm":28,"read_time":29},85920,7971461740886,"Theodore","https://ap-avatar.wpscdn.com/davatar_3d24733baf745e90a7e4bdd5f77d97b2",8,"Research & Report","GRASP Granularity-Aware Search Policy for Agentic RAG","Agentic retrieval-augmented generation (RAG) enables language models to iteratively reason by generating search queries, retrieving evidence, and predicting answers, yet it struggles with deciding when to retrieve, selecting lexical versus semantic retrieval, and controlling context granularity to avoid distracting tokens. GRASP introduces a reinforcement learning framework that adaptively coordinates complementary retrieval tools. It provides semantic search, keyword search, and paragraph-reading actions for sentence-level evidence and expandable context only when needed. Rewards jointly optimize answer accuracy, grounded reading, complementary search quality, and turn efficiency.","GRASP: GRanularity-Aware Search Policy for Agentic RAG  \nVarun Gandhi1,* , Jaewook Lee1,* Shantanu Todmal1 , Franck Dernoncourt2 , Ryan Rossi2 , Zichao Wang2 , Andrew Lan1 ,  \n1University of Massachusetts Amherst, 2Adobe Research  \n*Equal contribution.  \nCorrespondence: [vgandhi@umass.edu](vgandhi@umass.edu)  \narXiv :2607 . 10463v 1 [ cs .AI] 11 Jul 2026  \nAbstract  \nAgentic retrieval-augmented generation (RAG) extends static RAG by allowing language models to iteratively reason, generate search queries, retrieve evidence, and predict answers. However, it remains challenging for models to decide when to retrieve, whether to use lexical matching or semantic similarity, and how to control context granularity to prevent irrelevant tokens from interfering with agent reasoning. In this paper, we introduce GRASP, a reinforcement learning (RL) framework for training agents to adaptively coordinate complementary retrieval tools during multi-step reasoning. GRASP provides the agent with semantic search, keyword search, and paragraph-reading actions, enabling it to retrieve sentence-level evidence and expand further context only when needed. We train the policy with a reward that jointly accounts for answer accuracy, grounded reading, complementary search, and turn efficiency. Experiments on multi-hop reasoning benchmarks show that GRASP improves both retrieval recall and downstream question answering performance compared with singlestep retrieval, prompting-based agentic RAG, and RL-based retrieval baselines. Qualitative and ablation analyses show that the learned policy develops interpretable skimming and scanning behavior: it uses semantic search for broad exploration, paragraph reading for local verification, and keyword search for entity-specific evidence. These results suggest that learning to coordinate retrieval signals and context granularity is critical for agent’s correct reasoning.  \n1 Introduction  \nLarge language models (LLMs) have demonstrated strong capabilities in understanding and generating natural language, enabling document-centered tasks such as long document summarization, information extraction, and reasoning for question answering (Perot et al., 2024 ; Jiao et al., 2023 ; Liet al., 2024) . However, their parametric knowl-  \nedge is largely limited to information acquired during pretraining, which can lead to factual inaccuracies and limited adaptability when up-todate or domain-specific knowledge is required (Liet al., 2025a ; Wang et al., 2024 ; Nie et al., 2025) . Retrieval-Augmented Generation (RAG) has emerged as a mainstream approach to addressing this limitation by incorporating external knowledge at inference time.  \nTraditional, static RAG is typically formulated as a static, single-step pipeline: A retriever first identifies relevant chunks from external documents, then an LLM generates a response conditioned on the retrieved evidence. Here, a chunk refers to a smaller textual unit such as a passage, paragraph, or fixed-length span, obtained by segmenting a long document, serving as the basic unit of retrieval. The retriever ranks chunks according to their relevance to the input query, using either lexical matching methods, such as BM25, or semantic similarity methods, such as dense retrieval models that encode queries and passages into continuous vector representations (Karpukhin et al., 2020) . In practice, the initially retrieved chunks are often further refined by a re-ranker, such as a cross-encoder, which computes relevance scores between query and chunk to improve final retrieval results (Renet al., 2021) .  \nRecently, RAG has evolved from static pipelines into agentic RAG, where the LLM acts as an agent that iteratively reasons, generates queries, retrieves evidence, and synthesizes answers over multiple steps. This evolution changes the what the retriever does from simply ranking a fixed set of candidate chunks to deciding how retrieval should happen at each step. In particular, agentic RAG int","cbCaik0LqKTHkAoJ","https://ap.wps.com/l/cbCaik0LqKTHkAoJ","pdf",1231621,4,1,27,"English","en",105,"# Abstract\n# 1 Introduction\n## Background: Static RAG and retrieval units\n## From static to agentic RAG\n## Key challenges: signals, granularity, and tool coordination\n## Motivation: context control for multi-hop reasoning","[{\"question\":\"What problem does GRASP target in agentic RAG?\",\"answer\":\"GRASP targets difficulties in deciding when to retrieve, which retrieval mode to use (lexical vs. semantic), and how to control context granularity so irrelevant tokens do not interfere with multi-step reasoning.\"},{\"question\":\"How does GRASP coordinate retrieval across multiple reasoning steps?\",\"answer\":\"GRASP uses a reinforcement learning policy that provides complementary retrieval tools—semantic search, keyword search, and paragraph reading—allowing the agent to retrieve sentence-level evidence and expand context only when necessary.\"},{\"question\":\"What evidence does GRASP claim from experiments?\",\"answer\":\"Experiments on multi-hop reasoning benchmarks show improved retrieval recall and downstream question answering versus single-step retrieval, prompting-based agentic RAG, and RL-based retrieval baselines, supported by qualitative and ablation analyses of learned skimming and scanning behavior.\"}]",1784207172,68,{"code":4,"msg":31,"data":32},"ok",{"site_id":25,"language":24,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"grasp-granularity-aware-search-policy-for-agentic-rag","",{"@graph":36,"@context":85},[37,53,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":21},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":50},"https://docshare.wps.com/document/research-report/",3,{"item":52,"name":13,"@type":43,"position":20},"https://docshare.wps.com/document/grasp-granularity-aware-search-policy-for-agentic-rag/85920/",{"url":52,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":24,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":41,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-24","2026-07-16",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What problem does GRASP target in agentic RAG?","Question",{"text":75,"@type":76},"GRASP targets difficulties in deciding when to retrieve, which retrieval mode to use (lexical vs. semantic), and how to control context granularity so irrelevant tokens do not interfere with multi-step reasoning.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does GRASP coordinate retrieval across multiple reasoning steps?",{"text":80,"@type":76},"GRASP uses a reinforcement learning policy that provides complementary retrieval tools—semantic search, keyword search, and paragraph reading—allowing the agent to retrieve sentence-level evidence and expand context only when necessary.",{"name":82,"@type":73,"acceptedAnswer":83},"What evidence does GRASP claim from experiments?",{"text":84,"@type":76},"Experiments on multi-hop reasoning benchmarks show improved retrieval recall and downstream question answering versus single-step retrieval, prompting-based agentic RAG, and RL-based retrieval baselines, supported by qualitative and ablation analyses of learned skimming and scanning behavior.","https://schema.org",{"og:url":52,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":52},"index,follow",{"doc_id":7,"site_id":25},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,135],{"id":21,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":20,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":46,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":46,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":46,"category_name":137,"show_sort_weight":106,"slug":138},19,"General","general"]