[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-83205-en":3,"doc-seo-83205-105":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},83205,1374391974468,"Eden","https://ap-avatar.wpscdn.com/davatar_29158cc5080c5b710cf443261637dec0",8,"Research & Report","Why Fake Unveiling the Semantic Vocabulary of Deepfake Detectors","Deepfake technology threatens information integrity, making robust detection and credible justification essential for real-world settings such as legal proceedings. Existing deepfake detectors often output only real/fake labels, while prior explainable approaches may require human annotations for artifact localization or produce text that lacks reliable evidence grounding. This work applies post-hoc explainable AI to analyze black-box detectors using Encoding-Decoding Direction Pairs (EDDP) to reveal the detectors’ implicit real/fake features, enabling global understanding, concept-level spatial localization, and counterfactual what-if analysis.","Why Fake ? Unveiling the Semantic Vocabulary of Deepfake Detectors  \nVazgken Vanian* [vvanian@iti.gr](vvanian@iti.gr)  \nAlexandros Doumanoglou* [aldoum@iti.gr](aldoum@iti.gr)  \nDimitris Zarpalas [zarpalas@iti.gr](zarpalas@iti.gr)  \nInformation Technologies Institute (ITI) Centre For Research and Technology HELLAS (CERTH)  \narXiv :2607 .072 16v 1 [ cs .CV] 8 Jul 2026  \nAbstract  \nDeepfake (DF) technology poses a significant threat to information integrity, driving the need for robust detection methods. Most DF detectors only consider predicting a binary label for whether the input is real or fake, lacking the justification required for real-world applications like legal proceedings. Explainable DF Detection has emerged to address this limitation, but existing techniques frequently fall short by either relying on human annotations for precise artifact localization or generating superficially plausible textual explanations without grounding. This work investigates the use of post-hoc explainable AI (XAI) to analyze the decision-making process of state-of-the-art black-box DF detectors. Specifically, we employ Encoding-Decoding Direction Pairs (EDDP), a technique suitable for uncovering the concept space of DF detectors (their semantic vocabulary) as well as the mechanism for writing and reading concept information to and from internal representations. Our analysis reveals previously hidden real and fake features learned implicitly during detector training, offering nuanced explanations unattainable through conventional methods. This enables global model understanding, spatially aware concept localization, and counterfactual what-if analysis, all contributing to a deeper comprehension of DF detection strategies.  \n1. Introduction  \nThe proliferation of deepfake (DF) content presents an escalating challenge to information integrity across various domains, from journalism and politics to legal proceedings [7, 17] . As generative models continue to advance, realism has surpassed previously unimaginable thresholds, necessitating the development of robust detection methodologies. The majority of existing approaches focus on binary classification [1, 34] . While effective, this approach suffers from a critical limitation: a lack of transparency and justification.  \n*Equal contribution  \nIn contexts demanding accountability a simple real or fake label is insufficient. Instead, understanding why a piece of content is flagged as manipulated is paramount.  \nThis need for explainability has spurred recent research into Explainable Deepfake Detection (XDFD). Existing approaches broadly fall into two categories: spatio-temporal localization [5, 15, 27] and textual explanation methods [14] .  \nSpatio-temporal localization methods aim to identify manipulated regions by highlighting areas of the image or segments of the video that are presumed to contain forged content. The motivation is to move beyond binary prediction and provide visual evidence supporting the decision. However, in practice, these methods often produce coarse localization, frequently defaulting to highlighting the entire face rather than isolating specific manipulation artifacts. As a result, they provide limited insight into what precise cues led to the prediction.  \nTextual explanation methods instead generate natural language justifications describing why an image was classified as fake. While these explanations improve human interpretability, they often lack accurate spatial grounding and may not faithfully reflect the underlying evidence used by the detector. Consequently, despite their explanatory intent, existing XDFD methods primarily function as prediction tools with attached justification mechanisms, rather than systems that deliver precise and causally grounded explanations.  \nIn this work, we explore XDFD from a different perspective by analyzing deepfake detectors through the lens of post-hoc Explainable Artificial Intelligence (XAI) . Rather than modifying detecto","cbCaifzmvt3bzamf","https://ap.wps.com/l/cbCaifzmvt3bzamf","pdf",634114,2,1,10,"English","en",105,"# Abstract\n# 1. Introduction\n## Explainable Deepfake Detection (XDFD)\n## Spatio-temporal localization vs textual explanations\n## Post-hoc concept-based XAI with EDDP\n# 2. Related Work\n## 2.1 DeepFake Generation","[{\"question\":\"Why are binary real/fake labels insufficient for deepfake detection in accountability contexts?\",\"answer\":\"In scenarios requiring justification, a simple real or fake label does not explain why content is flagged as manipulated. The document argues that understanding the underlying reason is crucial for real-world use such as legal proceedings.\"},{\"question\":\"What limitations do existing Explainable Deepfake Detection approaches have?\",\"answer\":\"Spatio-temporal localization often produces coarse results, frequently highlighting the entire face instead of specific manipulation cues. Textual explanation methods may lack accurate spatial grounding and may not faithfully reflect the evidence used by the detector.\"},{\"question\":\"How does the proposed method use EDDP for explainability?\",\"answer\":\"The approach uses Encoding-Decoding Direction Pairs (EDDP) to uncover the semantic vocabulary (concept space) of deepfake detectors and the mechanism for encoding/decoding concept information in internal representations. This supports global model understanding, concept-based spatial localization, and counterfactual what-if analysis.\"}]",1784185936,25,{"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},"why-fake-unveiling-the-semantic-vocabulary-of-deepfake-detectors","",{"@graph":36,"@context":85},[37,53,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,47,50],{"item":41,"name":42,"@type":43,"position":21},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":20},"https://docshare.wps.com/document/","Document",{"item":48,"name":12,"@type":43,"position":49},"https://docshare.wps.com/document/research-report/",3,{"item":51,"name":13,"@type":43,"position":52},"https://docshare.wps.com/document/why-fake-unveiling-the-semantic-vocabulary-of-deepfake-detectors/83205/",4,{"url":51,"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","2026-07-23","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},"Why are binary real/fake labels insufficient for deepfake detection in accountability contexts?","Question",{"text":75,"@type":76},"In scenarios requiring justification, a simple real or fake label does not explain why content is flagged as manipulated. The document argues that understanding the underlying reason is crucial for real-world use such as legal proceedings.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"What limitations do existing Explainable Deepfake Detection approaches have?",{"text":80,"@type":76},"Spatio-temporal localization often produces coarse results, frequently highlighting the entire face instead of specific manipulation cues. Textual explanation methods may lack accurate spatial grounding and may not faithfully reflect the evidence used by the detector.",{"name":82,"@type":73,"acceptedAnswer":83},"How does the proposed method use EDDP for explainability?",{"text":84,"@type":76},"The approach uses Encoding-Decoding Direction Pairs (EDDP) to uncover the semantic vocabulary (concept space) of deepfake detectors and the mechanism for encoding/decoding concept information in internal representations. This supports global model understanding, concept-based spatial localization, and counterfactual what-if analysis.","https://schema.org",{"og:url":51,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":51},"index,follow",{"doc_id":7,"site_id":25},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,134],{"id":21,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":20,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":52,"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":22,"doc_module":4,"doc_module_name":46,"category_name":132,"show_sort_weight":22,"slug":133},"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":46,"category_name":136,"show_sort_weight":106,"slug":137},19,"General","general"]