Exploiting Modality-Specific Features for Multi-Modal Manipulation Detection and Grounding
Open accessAbstract
AI-synthesized text and images have gained significant attention, particularly due to the widespread dissemination of multi-modal manipulations on the internet, which has resulted in numerous negative impacts on society. Existing methods for multi-modal manipulation detection and grounding primarily focus on fusing vision-language features to make predictions, while overlooking the importance of modality-specific features, leading to sub-optimal results. In this paper, we construct a simple and novel transformer-based framework for multi-modal manipulation detection and grounding tasks. Our framework simultaneously explores modality-specific features while preserving the capability for multi-modal alignment. To achieve this, we introduce visual/language pre-trained encoders and dual-branch cross-attention (DCA) to extract and fuse modality-unique features. Furthermore, we design decoupled fine-grained classifiers (DFC) to enhance modality-specific feature mining and mitigate modality competition. Moreover, we propose an implicit manipulation query (IMQ) that adaptively aggregates global contextual cues within each modality using learnable queries, thereby improving the discovery of forged details. Extensive experiments on the DGM4dataset demonstrate the superior performance of our proposed model compared to state-of-the-art approaches.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The performance of any deepfake detection method is highly dependent on the quantitative and qualitative aspects of the datasets used for development.
“The performance of any deepfake detection method is highly dependent on the quantitative and qualitative aspects of the datasets used for development [18, 21, 25, 40, 44, 47, 49, 51, 52, 61, 63, 69, 72, 75].”
From AV-Deepfake1M: A Large-Scale LLM-Driven Audio-Visual Deepfake Dataset (2024) · cites this paper as [61]
Wang et al. [241] emphasized exploiting modality-specific features for enhanced detection and grounding of multi-modal manipulations.
“Wang et al. [241] emphasized exploiting modalityspecific features for enhanced detection and grounding of multi-modal manipulations.”
From Evolving from Single-modal to Multi-modal Facial Deepfake Detection: Progress and Challenges (2024) · cites this paper as [241]
Wang et al.'s framework incorporates modality-specific fine-grained classifiers to improve manipulation classification accuracy.
“incorporates modalityspecific fine-grained classifiers to improve manipulation classification accuracy.”
From Evolving from Single-modal to Multi-modal Facial Deepfake Detection: Progress and Challenges (2024) · cites this paper as [241]
Wang et al.'s dual-branch cross-attention framework facilitates bi-directional interaction between text and visual modalities.
“Their dual-branch crossattention framework facilitates bi-directional interaction between text and visual modalities”
From Evolving from Single-modal to Multi-modal Facial Deepfake Detection: Progress and Challenges (2024) · cites this paper as [241]
Cites
End-to-End Object Detection with Transformers
Carion, N., Massa, F., Synnaeve, G., Usunier, N., Kirillov, A., & Zagoruyko, S. (2020) · Lecture Notes in Computer Science · doi:10.1007/978-3-030-58452-8_13 · Cited by 18,285 · see citing works
End-to-End Object Detection with Transformers
Carion, N., Massa, F., Synnaeve, G., Usunier, N., Kirillov, A., & Zagoruyko, S. (2020) · arXiv · doi:10.48550/arxiv.2005.12872 · Cited by 18,253 · cited 1×
“Drawing inspiration from the works of DETR [16] and MaskFormer [17], we propose an implicit manipulation query (IMQ) module, which consists of two components: I-IMQ and T-IMQ.”
MethodLearning Transferable Visual Models From Natural Language Supervision
Radford, A., Kim, J. W., Hallacy, C., Ramesh, A., Goh, G., Agarwal, S., Sastry, G., Askell, A., Mishkin, P., Clark, J., et al. (2021) · arXiv · doi:10.48550/arxiv.2103.00020 · Cited by 15,293 · cited 1× · see citing works
“Categories Binary Cls Multi-Label Cls Image Grounding Text Grounding Methods Params AUC EER ACC mAP CF1 OF1 IoUmean IoU50 IoU75 Precision Recall F1 CLIP [26] - 83.”
MethodGeneralized Intersection Over Union: A Metric and a Loss for Bounding Box Regression
Rezatofighi, H., Tsoi, N., Gwak, J., Sadeghian, A., Reid, I., & Savarese, S. (2019) · 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2019.00075 · Cited by 6,612 · cited 1× · see citing works
“The manipulated image grounding loss is calculated by combining the L1 loss and GIoU loss [18], where y mig represents the manipulated image grounding label: The f tm and t tok are dimensionally reduced, and an inner product is performed in the feature dimension to predict whether each token is…”
MethodPer-Pixel Classification is Not All You Need for Semantic SegmentationPreprint
Cheng, B., Schwing, A. G., & Kirillov, A. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2107.06278 · Cited by 1,048 · cited 1× · see citing works
“Drawing inspiration from the works of DETR [16] and MaskFormer [17], we propose an implicit manipulation query (IMQ) module, which consists of two components: I-IMQ and T-IMQ.”
MethodViLT: Vision-and-Language Transformer Without Convolution or Region Supervision
Kim, W., Son, B., & Kim, I. (2021) · arXiv · doi:10.48550/arxiv.2102.03334 · Cited by 951 · see citing works
Defending Against Neural Fake News
Zellers, R., Holtzman, A., Rashkin, H., Bisk, Y., Farhadi, A., Roesner, F., & Choi, Y. (2019) · arXiv · doi:10.48550/arxiv.1905.12616 · Cited by 569 · cited 1× · see citing works
“These deepfake products [1, 2] have the potential to spread widely on social media.”
BackgroundGLTR: Statistical Detection and Visualization of Generated Text
Gehrmann, S., Strobelt, H., & Rush, A. (2019) · Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics: System Demonstrations · doi:10.18653/v1/p19-3019 · Cited by 470 · cited 1× · see citing works
“Consequently, this threat has garnered significant attention in the fields of computer vision and natural language processing, leading to the proposal of various methods for detecting fake faces and AI-generated text [3, 4, 5, 6, 7, 8, 9, 10].”
BackgroundThe deepfake detection challenge (dfdc) dataset
Dolhansky, B., Bitton, J., Pflaum, B., Lu, J., Howes, R., Wang, M., & Ferrer, C. C. (2020) · arXiv · doi:10.48550/arxiv.2006.07397 · Cited by 427 · see citing works
An Empirical Study of Training End-to-End Vision-and-Language Transformers
Dou, Z.-Y., Xu, Y., Gan, Z., Wang, J., Wang, S., Wang, L., Zhu, C., Zhang, P., Yuan, L., Peng, N., et al. (2022) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52688.2022.01763 · Cited by 366 · cited 1× · see citing works
“The pre-training weights are derived from METER [19].”
MethodSelf-citationHuang, Y., Lin, J.-Y., Zhou, C., Yang, H.-X., & Huang, L.-B. (2022) · International Conference on Machine Learning · doi:10.48550/arxiv.2203.12221 · Cited by 238 · cited 1× · see citing works
“However, this approach may suffer from modality competition [15], where one modality is learned well while the other is not fully explored.”
BackgroundImplicit Identity Driven Deepfake Face Swapping Detection
Huang, B., Wang, Z., Yang, J., Ai, J., Zou, Q., Wang, Q., & Ye, D. (2023) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52729.2023.00436 · Cited by 189 · cited 1× · see citing works
“Consequently, this threat has garnered significant attention in the fields of computer vision and natural language processing, leading to the proposal of various methods for detecting fake faces and AI-generated text [3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14].”
BackgroundF <sup>2</sup> Trans: High-Frequency Fine-Grained Transformer for Face Forgery Detection
Miao, C., Tan, Z., Chu, Q., Liu, H., Hu, H., & Yu, N. (2023) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2022.3233774 · Cited by 172 · cited 1× · see citing works
“Consequently, this threat has garnered significant attention in the fields of computer vision and natural language processing, leading to the proposal of various methods for detecting fake faces and AI-generated text [3, 4, 5, 6, 7, 8, 9, 10].”
BackgroundSelf-citationBootstrapping Multi-View Representations for Fake News Detection
Ying, Q.-C., Hu, X., Zhou, Y., Qian, Z.-X., Zeng, D., & Ge, S.-M. (2022) · AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v37i4.25670 · Cited by 155 · cited 1× · see citing works
“However, these methods [15, 16, 17] are unable to determine specific types of manipulation or localize manipulation positions, lacking applicability and interpretability.”
BackgroundZhuang, W., Chu, Q., Tan, Z., Liu, Q., Yuan, H.-J., Miao, C.-T., Luo, Z., & Yu, N.-H. (2022) · European Conference on Computer Vision · doi:10.48550/arxiv.2210.12752 · Cited by 147 · cited 1× · see citing works
“Consequently, this threat has garnered significant attention in the fields of computer vision and natural language processing, leading to the proposal of various methods for detecting fake faces and AI-generated text [3, 4, 5, 6, 7, 8, 9, 10].”
BackgroundSelf-citationAltFreezing for More General Video Face Forgery Detection
Wang, Z., Bao, J., Zhou, W., Wang, W., & Li, H. (2023) · 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr52729.2023.00402 · Cited by 137 · cited 1× · see citing works
“Consequently, this threat has garnered significant attention in the fields of computer vision and natural language processing, leading to the proposal of various methods for detecting fake faces and AI-generated text [3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14].”
BackgroundHierarchical Frequency-Assisted Interactive Networks for Face Manipulation Detection
Miao, C.-T., Tan, Z., Chu, Q., Yu, N.-H., & Guo, G.-D. (2022) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2022.3198275 · Cited by 110 · cited 1× · see citing works
“Consequently, this threat has garnered significant attention in the fields of computer vision and natural language processing, leading to the proposal of various methods for detecting fake faces and AI-generated text [3, 4, 5, 6, 7, 8, 9, 10].”
BackgroundSelf-citationDetecting and Grounding Multi-Modal Media Manipulation
Shao, R., Wu, T., & Liu, Z. (2023) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52729.2023.00667 · Cited by 86 · cited 6× · see citing works
“In HAMMER [14], the embedding of the [CLS] token for binary classification of the four manipulation types.”
BackgroundMethodInfluentialTransformer-Based Feature Compensation and Aggregation for DeepFake Detection
Tan, Z., Yang, Z., Miao, C., & Guo, G. (2022) · IEEE Signal Processing Letters · doi:10.1109/lsp.2022.3214768 · Cited by 59 · cited 1× · see citing works
“Consequently, this threat has garnered significant attention in the fields of computer vision and natural language processing, leading to the proposal of various methods for detecting fake faces and AI-generated text [3, 4, 5, 6, 7, 8, 9, 10].”
BackgroundSelf-citationCross-Modal Knowledge Distillation in Multi-Modal Fake News Detection
Wei, Z., Pan, H., Qiao, L.-B., Niu, X., Dong, P., & Li, D. (2022) · IEEE International Conference on Acoustics, Speech, and Signal Processing · doi:10.1109/icassp43922.2022.9747280 · Cited by 56 · cited 1×
“Yet, in everyday life, multi-modal media content in the form of image-text pairs is more prevalent. of our approach.”
BackgroundLearning Forgery Region-Aware and ID-Independent Features for Face Manipulation Detection
Miao, C.-T., Chu, Q., Li, W.-H., Li, S., Tan, Z., Zhuang, W., & Yu, N.-H. (2022) · IEEE Transactions on Biometrics Behavior and Identity Science · doi:10.1109/tbiom.2021.3119403 · Cited by 48 · cited 1× · see citing works
“Consequently, this threat has garnered significant attention in the fields of computer vision and natural language processing, leading to the proposal of various methods for detecting fake faces and AI-generated text [3, 4, 5, 6, 7, 8, 9, 10].”
BackgroundSelf-citationFrequency-Aware Attentional Feature Fusion for Deepfake Detection
Tian, C., Luo, Z.-M., Shi, G., & Li, S.-Z. (2023) · IEEE International Conference on Acoustics, Speech, and Signal Processing · doi:10.1109/icassp49357.2023.10094654 · Cited by 39 · cited 1× · see citing works
“Consequently, this threat has garnered significant attention in the fields of computer vision and natural language processing, leading to the proposal of various methods for detecting fake faces and AI-generated text [3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14].”
BackgroundTowards Generalizable and Robust Face Manipulation Detection via Bag-of-feature
Miao, C., Chu, Q., Li, W., Gong, T., Zhuang, W., & Yu, N. (2021) · IEEE International Conference on Visual Communications and Image Processing (VCIP) · doi:10.1109/vcip53242.2021.9675331 · Cited by 35 · cited 1× · see citing works
“Consequently, this threat has garnered significant attention in the fields of computer vision and natural language processing, leading to the proposal of various methods for detecting fake faces and AI-generated text [3, 4, 5, 6, 7, 8, 9, 10].”
BackgroundSelf-citationLiu, H., Tan, Z., Chen, Q., Wei, Y., Zhao, Y., & Wang, J. (2024) · International Journal of Computer Vision · doi:10.1007/s11263-024-02245-x · Cited by 25 · cited 1× · see citing works
“UFAFormer [19] introduces the frequency domain as a complementary viewpoint, develops a forgery-aware mutual module to interact with image and frequency information, and proposes a unified decoder responsible for gathering modality-specific forgery information.”
BackgroundSelf-citationADT: Anti-Deepfake Transformer
Wang, P., Liu, K., Zhou, W.-B., Zhou, H., Liu, H., Zhang, W., & Yu, N.-H. (2022) · IEEE International Conference on Acoustics, Speech, and Signal Processing · doi:10.1109/icassp43922.2022.9746888 · Cited by 18 · cited 1× · see citing works
“Consequently, this threat has garnered significant attention in the fields of computer vision and natural language processing, leading to the proposal of various methods for detecting fake faces and AI-generated text [3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14].”
BackgroundSelf-citationZhuang, W., Chu, Q., Yuan, H.-J., Miao, C.-T., Liu, B., & Yu, N.-H. (2022) · IEEE International Conference on Multimedia and Expo · doi:10.1109/icme52920.2022.9859586 · Cited by 17 · cited 1× · see citing works
“Consequently, this threat has garnered significant attention in the fields of computer vision and natural language processing, leading to the proposal of various methods for detecting fake faces and AI-generated text [3, 4, 5, 6, 7, 8, 9, 10].”
BackgroundSelf-citationCross-Modal Adversarial Contrastive Learning for Multi-Modal Rumor Detection
Zou, T., Zhong, Q., Li, P., & Zhu, Q. (2023) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49357.2023.10096883 · Cited by 6 · cited 1× · see citing works
“Yet, in everyday life, multi-modal media content in the form of image-text pairs is more prevalent. of our approach.”
Background
Flagged first, then by citations
1 of the 28 references couldn’t be matched to a known paper.
Cited by
Sorted by citations
Detecting Multimedia Generated by Large AI Models: A SurveyPreprint
Lin, L., Gupta, N., Zhang, Y., Ren, H.-N., Liu, C.-H., Ding, F., Wang, X., Li, X., Verdoliva, L., & Hu, S. (2024) · arXiv.org · doi:10.48550/arxiv.2402.00045 · Cited by 117 · see citing works
“Wang et al. [146] construct a simple and novel transformer-based framework for the same task.”
AV-Deepfake1M: A Large-Scale LLM-Driven Audio-Visual Deepfake Dataset
Cai, Z., Ghosh, S., Adatia, A. P., Hayat, M., Dhall, A., Gedeon, T., & Stefanov, K. (2024) · Proceedings of the 32nd ACM International Conference on Multimedia · doi:10.1145/3664647.3680795 · Cited by 59 · see citing works
Multimodal Misinformation Detection by Learning from Synthetic Data with Multimodal LLMs
Zeng, F., Li, W., Gao, W., & Pang, Y. (2024) · Conference on Empirical Methods in Natural Language Processing · doi:10.48550/arxiv.2409.19656 · Cited by 27
“…al., 2021; Qian et al., 2021; Ghorbanpour et al., 2023; Zhou et al., 2023; Liu et al., 2023c; Chen et al., 2023), or from the same large-scale synthetic dataset (Aneja et al., 2023; Abdelnabi et al., 2022; Shao et al., 2023; Jia et al., 2023; Mu et al., 2023; Wang et al., 2024a; Qi et al., 2024).”
BackgroundAudio-Visual Multimodal Deepfake Detection Leveraging Emotional Recognition
Alsaeedi, A., Almansour, A., & Jamal, A. (2025) · International Journal of Advanced Computer Science and Applications · doi:10.14569/ijacsa.2025.0160622 · Cited by 12 · see citing works
“Wang et al. [28] achieved the highest result for text-visual models at 91.42%.”
InfluentialLi, Y., Yang, Y., Tan, Z., Liu, H., Chen, W., Zhou, X. Y., & Lei, Z. (2025) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52734.2025.00863 · Cited by 7 · see citing works
“This means that we no longer need to provide additional prediction head and supervision for grounding text manipulation as previous methods [43, 47].”
InfluentialEvolving from Single-modal to Multi-modal Facial Deepfake Detection: Progress and Challenges
Liu, P., Tao, Q., & Zhou, J. T. (2024) · arXiv · doi:10.48550/arxiv.2406.06965 · Cited by 6 · see citing works
“Wang et al. [241] emphasized exploiting modality-specific features for enhanced detection and grounding of multi-modal manipulations.”
Fusion in Context: A Multimodal Approach to Affective State Recognition
Mohamed, Y., Lemaignan, S., Guneysu, A., Jensfelt, P., & Smith, C. (2024) · IEEE International Symposium on Robot and Human Interactive Communication · doi:10.1109/ro-man63969.2025.11217904 · Cited by 5 · see citing works
“This study utilizes a dataset collected by [25], which captures participants’ affective states during a tangible Pacman game designed with multiple configurations and can induce four affects: frustration, enjoyment, boredom and neutral [25], [9].”
Review of AIGC passive detection and proactive protection
Liu, X.-L., Liu, H.-C., Zhao, Y., Ni, R.-R., Li, X.-L., & Guo, M.-Z. (2025) · Scientia Sinica Informationis · doi:10.1360/ssi-2025-0043 · Cited by 3 · see citing works
“Wang 等 [225] 提出了一种基于 变换器的多模态操控检测与定位框架 , 通过双分支交叉注意力和隐式操控查询显著提升模态特征挖掘 和伪造细节发现能力 .”
Qin, J., Chen, J., Chen, T., & Yang, Z. (2026) · Expert Systems with Applications · doi:10.1016/j.eswa.2026.132076 · Cited by 1 · see citing works
Lee, H.-W., Zhang, M.-L., Lee, H.-S., & Lee, C.-S. (2025) · 2025 4th International Symposium on Computer Applications and Information Technology (ISCAIT) · doi:10.1109/iscait64916.2025.11010515 · Cited by 2
BASNet: Boundary Assisted Network for Image Splicing Forgery Detection
Liang, E.-J., Zhang, K.-Y., Hua, Z.-Y., & Jia, X. (2025) · IEEE transactions on multimedia · doi:10.1109/tmm.2025.3604911 · Cited by 2 · see citing works
“Some researchers have explored fine-grained classifi-cation for image forgery region detection [18], [19], [20].”
Evidence-Grounded Forensic Reasoning for Detecting and Grounding Multi-Modal Media ManipulationPreprint
Yeh, Y., Li, Y., Hu, X., Lei, Z., & Yang, Y. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2608.08009
CIEC: Coupling Implicit and Explicit Cues for Multimodal Weakly Supervised Manipulation Localization
Yu, X.-Q., Lu, W., Luo, X.-Y., & Yang, R. (2026) · arXiv.org · doi:10.48550/arxiv.2602.02175 · Cited by 1
Adaptive Contribution Modulation For Multi-Modal Manipulation Media Detection and Grounding
Li, Y., & Leng, B. (2025) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49660.2025.10890433
“…al. [27] focuses on performing inference tasks by integrating and aligning mult-imodal features from the perspective of task inference, Wang et al. [28] proposed a dual branch cross attention framework that focuses on modality specific features of fine-grained classifier interactions from the…”
Dual-Stream Decoupling of Sentiment and Facts for Multi-Modal Forgery Detection
Zhou, Y., Wan, H., Wu, K., Zhu, Y., Li, S., & Zhang, J. (2026) · International Conference on Big Data and Information Analytics (BigDIA) · doi:10.1109/bigdia71496.2026.11684704
“): Takes the text fact feature as the query, and the visual context feature as the key and value for cross-attention calculation[14], and adds a residual connection to retain the original feature information: (10) This path is specifically designed to detect logical factual contradictions caused by…”
DAIN: Dynamic Agent-Based Interaction Network for Efficient and Collaborative Multimodal Reasoning
Chen, X., Li, Y., Wang, Z., Zhang, H.-Y., Liu, R., & Zhao, M. (2026) · arXiv.org · doi:10.48550/arxiv.2606.30189
“Prior approaches have focused on attention visualization [4, 3] or individual modality attribution [24, 21, 50].”
Bridging Pixels and Words: Mask-Aware Local Semantic Fusion for Multimodal Media Verification
Chen, Z.-Z., Wei, P., Ren, Z.-Y., Li, H., & Yin, X. (2026) · Computer Vision and Pattern Recognition · doi:10.48550/arxiv.2603.26052
“Inspired by [33, 44], we propose a dual-path semantic extraction framework comprising three components: (1 For the global feature, we define it as where v clsimg is the result of v passing through global average pooling.”
Influential
- Authors
- Wang, J., Liu, B., Miao, C., Zhao, Z., Zhuang, W., Chu, Q., & Yu, N.
- Affiliations
- University of Science and Technology of China (CN)
- Venue
- ICASSP 2024 - 2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)
- Publisher
- IEEE
- Published
- 2024-04-14
- Pages
- 4935-4939
- ISSN
- 1520-6149 (ISSN-L)
- OpenAlex
- W4392903163
- Cited by
- 12
- References
- 28
- Citations by year
- 2026: 4 · 2025: 8
- FWCI
- 2.49
- Type
- conference paper
- Language
- en
- Open access
- green
- arxiv.org
- Locations
- IEEE International Conference on Acoustics Speech and Signal Processing (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- Funders
- National Natural Science Foundation of China
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Digital Media Forensic Detection
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Multimodal Machine Learning Applications
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Generative Adversarial Networks and Image Synthesis
- Keywords
- modality-specific features, multimodal alignment
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
