[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"detail-sidebar-cat-0-en-105":3,"doc-seo-327287-105":59,"doc-detail-327287-en":131},{"code":4,"msg":5,"data":6},0,"success",[7,13,18,23,28,33,38,43,48,51,55],{"id":8,"doc_module":4,"doc_module_name":9,"category_name":10,"show_sort_weight":11,"slug":12},1,"Document","Story & Novel",90,"story-novel",{"id":14,"doc_module":4,"doc_module_name":9,"category_name":15,"show_sort_weight":16,"slug":17},2,"Literature",80,"literature",{"id":19,"doc_module":4,"doc_module_name":9,"category_name":20,"show_sort_weight":21,"slug":22},4,"Exam",70,"exam",{"id":24,"doc_module":4,"doc_module_name":9,"category_name":25,"show_sort_weight":26,"slug":27},5,"Comic",60,"comic",{"id":29,"doc_module":4,"doc_module_name":9,"category_name":30,"show_sort_weight":31,"slug":32},6,"Technology",50,"technology",{"id":34,"doc_module":4,"doc_module_name":9,"category_name":35,"show_sort_weight":36,"slug":37},7,"Healthcare",40,"healthcare",{"id":39,"doc_module":4,"doc_module_name":9,"category_name":40,"show_sort_weight":41,"slug":42},8,"Research & Report",30,"research-report",{"id":44,"doc_module":4,"doc_module_name":9,"category_name":45,"show_sort_weight":46,"slug":47},9,"Religion & Spirituality",20,"religion-spirituality",{"id":46,"doc_module":4,"doc_module_name":9,"category_name":49,"show_sort_weight":46,"slug":50},"World Cup","world-cup",{"id":52,"doc_module":4,"doc_module_name":9,"category_name":53,"show_sort_weight":52,"slug":54},10,"Lifestyle","lifestyle",{"id":56,"doc_module":4,"doc_module_name":9,"category_name":57,"show_sort_weight":24,"slug":58},19,"General","general",{"code":4,"msg":60,"data":61},"ok",{"site_id":62,"language":63,"slug":64,"title":65,"keywords":66,"description":67,"schema_data":68,"social_meta":124,"head_meta":126,"extra_data":128,"updated_unix":130},105,"en","multattnattrib-training-free-multimodal-attribution-in-long-document-question-answering","MULTATTNATTRIB - Training-Free Multimodal Attribution in Long Document Question Answering","","As grounded QA systems are increasingly deployed in AI assistants, accurately attributing generated answers to evidence is critical for user trust and model safety. While unimodal attributions have been explored in depth, the multimodal setting remains relatively under-researched. This work introduces MULTATTNATTRIB, a training-free method that uses a model’s prefill pass, selected attention heads, and calibrated thresholds to locate source evidence in long documents. It also presents MULTATTREVAL, a benchmark with fine-grained multimodal ground-truth annotations for answer components. Experiments show MULTATTNATTRIB consistently outperforms attribution-generation baselines and matches frontier models while producing attributions with much lower latency.",{"@graph":69,"@context":123},[70,84,106],{"@type":71,"itemListElement":72},"BreadcrumbList",[73,77,79,82],{"item":74,"name":75,"@type":76,"position":8},"https://docshare.wps.com","Home","ListItem",{"item":78,"name":9,"@type":76,"position":14},"https://docshare.wps.com/document/",{"item":80,"name":40,"@type":76,"position":81},"https://docshare.wps.com/document/research-report/",3,{"item":83,"name":65,"@type":76,"position":19},"https://docshare.wps.com/document/multattnattrib-training-free-multimodal-attribution-in-long-document-question-answering/327287/",{"url":83,"name":65,"@type":85,"image":86,"author":91,"headline":65,"publisher":94,"fileFormat":97,"inLanguage":63,"description":67,"dateModified":98,"datePublished":99,"encodingFormat":97,"isAccessibleForFree":100,"interactionStatistic":101},"DigitalDocument",{"url":87,"@type":88,"width":89,"height":90},"https://docshare.wps.com/thumbnails/multattnattrib-training-free-multimodal-attribution-in-long-document-question-answering/327287.png","ImageObject",300,407,{"name":92,"@type":93},"Olivia Brown","Person",{"url":74,"name":95,"@type":96},"DocShare","Organization","application/pdf","2026-11-03","2026-09-21",true,{"@type":102,"interactionType":103,"userInteractionCount":105},"InteractionCounter",{"@type":104},"ViewAction",16,{"@type":107,"mainEntity":108},"FAQPage",[109,115,119],{"name":110,"@type":111,"acceptedAnswer":112},"What is MULTATTNATTRIB designed to do?","Question",{"text":113,"@type":114},"MULTATTNATTRIB generates fine-grained attributions for answer components in long documents by locating supporting evidence across modalities using a training-free approach.","Answer",{"name":116,"@type":111,"acceptedAnswer":117},"How does MULTATTNATTRIB avoid iterative prompting during attribution?",{"text":118,"@type":114},"It extracts attribution signals directly from the model’s prefill pass, aggregates attention from selected retrieval heads, and applies a lightweight calibration procedure in a single inference pass.",{"name":120,"@type":111,"acceptedAnswer":121},"What is MULTATTREVAL and why is it introduced?",{"text":122,"@type":114},"MULTATTREVAL is a complementary benchmark dataset with fine-grained ground-truth multimodal attributions across five domains, created to evaluate multimodal attribution in long-form documents.","https://schema.org",{"og:url":83,"og:type":125,"og:title":65,"og:site_name":95,"og:description":67},"article",{"robots":127,"canonical":83},"index,follow",{"doc_id":129,"site_id":62},327287,1790471121,{"code":4,"msg":5,"data":132},{"doc_id":129,"user_id":133,"nickname":92,"user_avatar":134,"doc_module":4,"category_id":39,"category_name":40,"doc_title":65,"doc_description":67,"doc_content":135,"file_id":136,"file_url":137,"file_type":138,"file_size":139,"view_count":105,"is_deleted":4,"is_public":8,"is_downloadable":8,"audit_status":8,"page_count":140,"language":141,"language_code":63,"site_id":62,"html_lang":63,"table_of_contents":142,"faqs":143,"seo_title":144,"seo_description":67,"update_tm":145,"read_time":146,"preview_page_count":4},16904993612988,"https://ap-avatar.wpscdn.com/davatar_a8503ba1806abce46bf441b54a3ca4cd","MULTATTNATTRIB: Training-Free Multimodal Attribution  \nin Long Document Question Answering  \nDang Quang Thien Tran 1 * Quang V. Dang 1∗ Vinamra Tyagi 1∗ Sai Soorya Rao Veeravalli 1∗ Trang Nguyen 1 Ryan A. Rossi2  \nFranck Dernoncourt2 Nedim Lipka2 Koustava Goswami2 Samyadeep Basu2  \n1University of Massachusetts, Amherst 2Adobe Research, San Jose  \narXiv :2607 .0 1420v 3 [ cs .CL] 8 Jul 2026  \nAbstract  \nAs grounded QA systems are increasingly deployed in AI assistants, accurately attributing generated answers to evidence is critical for user trust and model safety. While unimodal attributions have been explored in depth, the multimodal setting remains relatively under-researched. As a result, we introduce MULTATTNATTRIB, a training-free attributiongeneration method that leverages a model’sprefill pass, selected attention heads, and calibrated thresholds to locate source evidence within a document. To establish baseline results for the method, we introduce MULTATTREVAL, a complementary benchmark dataset annotated with fine-grained, ground-truth attributions for answer components grounded in multimodal source documents. To our knowledge, this is the first evaluation dataset designed specifically for multimodal attribution in long-form documents. Experimental results show that MULTATTNATTRIB consistently outperforms a variety of attribution-generation methods, including several strong prompting-based approaches and matches the latest frontier models such as GPT 5.4 . Our method not only substantially improves attribution accuracy for both unimodal and multimodal attribution types, but also produces attributions at up to one-seventh of the direct inference latency compared to prompting on the same base model.  \n1 Introduction  \nBuilding user trust in AI systems is critical to the success of agentic workflows in both enterprise and consumer environments. In many settings, users cannot safely act on a generated answer without verifying its source and validity—even modern generative systems fully support fewer than 52% of their generated statements with accurate citations (Liu et al., 2023) . As a result, model grounding via attributions—localizing each answer component to its supporting evidence—has emerged as  \n* Equal contribution.  \na fundamental requirement for model deployment, particularly in domains such as medicine where ungrounded or hallucinated answers can have real negative impacts (Kim et al., 2025) .  \nThere have been increasing efforts to use attributions to ground document question-answer pairs, though most focus on text-only or otherwise unimodal settings. Current approaches typically rely on citation-style generation (Bohnet et al., 2022 ; Gao et al., 2023b ; Berchansky et al., 2024), retrieval-head or circuit isolation (Basuet al., 2025), or decomposition-based attribution methods (Ramu et al., 2024 ; Balasubramanianet al., 2026), which have thus far been explored only for text. Real documents, however, interleave text with images, charts, and other raster content. A robust attribution system must therefore identify not only the correct source, but also the supporting modality or combination of modalities.  \nThe multimodal long-document setting remains comparatively nascent, with preexisting approaches largely framing attribution as citation selection from pre-retrieved passages or images rather than as fine-grained localization within a single full-length document (Ma et al., 2025 ; Qi et al., 2024 ; Song et al., 2026) . A unique set of challenges arises in multimodal attribution that doesn’t apply in the unimodal case: determining both the correct modality (or modalities) and the correct source within it. Resolving how text and images jointly support a single attribution remains an open problem with significant downstream potential.  \nTo address this critical challenge, we propose MULTATTNATTRIB, a training-free multimodal attribution method that leverages attention patterns from a model’s prefill pass to localiz","cbCaiewxyJzwNC5V","https://ap.wps.com/l/cbCaiewxyJzwNC5V","pdf",938483,25,"English","# Abstract\n# 1 Introduction\n## User trust and grounding via attributions\n## Prior work and the unimodal focus\n## Challenges in multimodal long-document attribution\n## Proposed method: MULTATTNATTRIB\n## New benchmark: MULTATTREVAL\n## Results and contributions","[{\"question\":\"What is MULTATTNATTRIB designed to do?\",\"answer\":\"MULTATTNATTRIB generates fine-grained attributions for answer components in long documents by locating supporting evidence across modalities using a training-free approach.\"},{\"question\":\"How does MULTATTNATTRIB avoid iterative prompting during attribution?\",\"answer\":\"It extracts attribution signals directly from the model’s prefill pass, aggregates attention from selected retrieval heads, and applies a lightweight calibration procedure in a single inference pass.\"},{\"question\":\"What is MULTATTREVAL and why is it introduced?\",\"answer\":\"MULTATTREVAL is a complementary benchmark dataset with fine-grained ground-truth multimodal attributions across five domains, created to evaluate multimodal attribution in long-form documents.\"}]","MULTATTNATTRIB - Training-Free Multimodal Attribution in Long Document Question Answering | PDF",1789976910,63]