HalluciDoctor: Mitigating Hallucinatory Toxicity in Visual Instruction Data
Open accessAbstract
Multi-modal Large Language Models (MLLMs) tuned on machine-generated instruction-following data have demonstrated remarkable performance in various multi-modal understanding and generation tasks. However, the hallucinations inherent in machine-generated data, which could lead to hallucinatory outputs in MLLMs, remain under-explored. This work aims to investigate various hallucinations (i.e., object, relation, attribute hallucinations) and mitigate those hallucinatory toxicities in large-scale machine-generated visual instruction datasets. Drawing on the human ability to identify factual errors, we present a novel hallucination detection and elimination framework, HalluciDoctor, based on the cross-checking paradigm. We use our framework to identify and eliminate hallucinations in the training data automatically. Interestingly, HalluciDoctor also indicates that spurious correlations arising from long-tail object cooccurrences contribute to hallucinations. Based on that, we execute counterfactual visual instruction expansion to balance data distribution, thereby enhancing MLLMs' resistance to hallucinations. Comprehensive experiments on hallucination evaluation benchmarks show that our method successfully mitigates 44.6% hallucinations relatively and maintains competitive performance compared to LLaVA. The data and code for this paper are publicly available.11https://github.com/Yuqifan1117/HalluciDoctor
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?
LLMs have limited robustness and increased hallucination rates in low-probability contexts, where reliance on surface-level statistical correlations is insufficient.
“This interpretation is reinforced by studies that have shown LLMs’ limited robustness and increased hallucination rates in low-probability contexts, where reliance on surface-level statistical correlations proves insufficient19,54,58.”
From Limitations of large language models in clinical problem-solving arising from inflexible reasoning (2025) · cites this paper as [58]
Cites
Visual Genome: Connecting Language and Vision Using Crowdsourced Dense Image Annotations
Krishna, R., Zhu, Y., Groth, O., Johnson, J., Hata, K., Kravitz, J., Chen, S., Kalantidis, Y., Li, L.-J., Shamma, D. A., et al. (2017) · International Journal of Computer Vision · doi:10.1007/s11263-016-0981-7 · Cited by 5,107 · cited 1× · see citing works
“Following previous works [15, 20], we randomly select 500 unique images from the intersection of MSCOCO [19] and Visual Genome [9] for a more detailed evaluation.”
Chain-of-Thought Prompting Elicits Reasoning in Large Language Models
Wei, J., Wang, X., Schuurmans, D., Bosma, M., Ichter, B., Xia, F., Chi, E., Le, Q., & Zhou, D. (2022) · arXiv · doi:10.48550/arxiv.2201.11903 · Cited by 4,263 · see citing works
Training language models to follow instructions with human feedback
Ouyang, L., Wu, J., Jiang, X., Almeida, D., Wainwright, C. L., Mishkin, P., Zhang, C., Agarwal, S., Slama, K., Ray, A., et al. (2022) · arXiv · doi:10.48550/arxiv.2203.02155 · Cited by 4,208 · see citing works
InstructBLIP: Towards General-purpose Vision-Language Models with Instruction TuningPreprint
Dai, W., Li, J., Li, D., Tiong, A., Zhao, J.-Q., Wang, W., Li, B. A., Fung, P., & Hoi, S. (2023) · Neural Information Processing Systems · doi:10.48550/arxiv.2305.06500 · Cited by 4,191 · see citing works
Self-citationLlama 2: Open Foundation and Fine-Tuned Chat Models
Touvron, H., Martin, L., Stone, K., Albert, P., Almahairi, A., Babaei, Y., Bashlykov, N., Batra, S., Bhargava, P., Bhosale, S., et al. (2023) · arXiv · doi:10.48550/arxiv.2307.09288 · Cited by 2,604 · see citing works
Microsoft COCO: Common Objects in Context
Lin, T.-Y., Maire, M., Belongie, S., Bourdev, L., Girshick, R., Hays, J., Perona, P., Ramanan, D., Zitnick, C. L., & Dollár, P. (2014) · arXiv · doi:10.48550/arxiv.1405.0312 · Cited by 2,579 · see citing works
Liu, H., Li, C., Wu, Q., & Lee, Y. J. (2023) · Advances in Neural Information Processing Systems 36 · doi:10.52202/075280-1516 · Cited by 2,428 · see citing works
A Survey of Large Language Models
Zhao, W. X., Zhou, K., Li, J., Tang, T., Dong, Z., Hou, Y., Zhang, B., Min, Y., Zhang, J., Liu, P., et al. (2026) · Frontiers of Computer Science · doi:10.1007/s11704-026-60308-3 · Cited by 2,093 · cited 1× · see citing works
“With the remarkable generalizability of LLMs in a zero-shot setting [26, 33, 34, 42], early works integrating LLMs with visual modality have demonstrated impressive visual-language understanding ability [8, 11, 13, 31, 39].”
Self-citationA Survey of Large Language Models
Zhao, W. X., Zhou, K., Li, J., Tang, T., Wang, X., Hou, Y., Min, Y., Zhang, B., Zhang, J., Dong, Z., et al. (2023) · arXiv · doi:10.48550/arxiv.2303.18223 · Cited by 2,090
Self-citationFlamingo: a Visual Language Model for Few-Shot Learning
Alayrac, J.-B., Donahue, J., Luc, P., Miech, A., Barr, I., Hasson, Y., Lenc, K., Mensch, A., Millican, K., Reynolds, M., et al. (2022) · arXiv · doi:10.48550/arxiv.2204.14198 · Cited by 1,555 · see citing works
mPLUG-Owl: Modularization Empowers Large Language Models with MultimodalityPreprint
Ye, Q.-H., Xu, H., Xu, G., Ye, J., Yan, M., Zhou, Y., Wang, J., Hu, A., Shi, P.-C., Shi, Y., et al. (2023) · arXiv.org · doi:10.48550/arxiv.2304.14178 · Cited by 1,318 · see citing works
Self-Instruct: Aligning Language Models with Self-Generated Instructions
Wang, Y., Kordi, Y., Mishra, S., Liu, A., Smith, N. A., Khashabi, D., & Hajishirzi, H. (2023) · Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · doi:10.18653/v1/2023.acl-long.754 · Cited by 1,243 · see citing works
Li, J., Li, D., Savarese, S., & Hoi, S. (2023) · arXiv · doi:10.48550/arxiv.2301.12597 · Cited by 908 · see citing works
Self-citationA survey on multimodal large language models
Yin, S., Fu, C., Zhao, S., Li, K., Sun, X., Xu, T., & Chen, E. (2024) · National Science Review · doi:10.1093/nsr/nwae403 · Cited by 893 · see citing works
Liu, H., Li, C., Wu, Q., & Lee, Y. J. (2023) · arXiv · doi:10.48550/arxiv.2304.08485 · Cited by 682 · see citing works
Evaluating Object Hallucination in Large Vision-Language Models
Li, Y., Du, Y., Zhou, K., Wang, J., Xin, Z., & Wen, J.-R. (2023) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/2023.emnlp-main.20 · Cited by 580 · see citing works
MiniGPT-4: Enhancing Vision-Language Understanding with Advanced Large Language Models
Zhu, D., Chen, J., Shen, X., Li, X., & Elhoseiny, M. (2023) · arXiv · doi:10.48550/arxiv.2304.10592 · Cited by 469 · see citing works
Object Hallucination in Image Captioning
Rohrbach, A., Hendricks, L. A., Burns, K., Darrell, T., & Saenko, K. (2018) · Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing · doi:10.18653/v1/d18-1437 · Cited by 420 · see citing works
Analyzing and Mitigating Object Hallucination in Large Vision-Language ModelsPreprint
Zhou, Y.-Y., Cui, C.-H., Yoon, J., Zhang, L.-J., Deng, Z., Finn, C., Bansal, M., & Yao, H. (2023) · International Conference on Learning Representations · doi:10.48550/arxiv.2310.00754 · Cited by 399 · see citing works
Grounding DINO: Marrying DINO with Grounded Pre-Training for Open-Set Object DetectionPreprint
Liu, S., Zeng, Z., Ren, T., Li, F., Zhang, H., Yang, J., Jiang, Q., Li, C., Yang, J., Su, H., et al. (2023) · arXiv (Cornell University) · doi:10.48550/arxiv.2303.05499 · Cited by 247 · see citing works
Woodpecker: hallucination correction for multimodal large language models
Yin, S., Fu, C., Zhao, S., Xu, T. B., Wang, H., Sui, D., Shen, Y., Li, K., Sun, X., & Chen, E. (2024) · Science China Information Sciences · doi:10.1007/s11432-024-4251-x · Cited by 196 · see citing works
Peng, B., Li, C., He, P., Galley, M., & Gao, J. (2023) · arXiv · doi:10.48550/arxiv.2304.03277 · Cited by 183 · see citing works
Guo, J., Li, J., Li, D., Tiong, A. M. H., Li, B., Tao, D., & Hoi, S. C. H. (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/cvpr52729.2023.01046 · Cited by 164 · see citing works
Self-citationDetecting and Preventing Hallucinations in Large Vision Language Models
Gunjal, A., Yin, J., & Bas, E. (2024) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v38i16.29771 · Cited by 161 · see citing works
MME: A Comprehensive Evaluation Benchmark for Multimodal Large Language Models
Fu, C., Chen, P., Shen, Y., Qin, Y., Zhang, M., Xu, L., Yang, J., Zheng, X., Li, K., Sun, X. W., et al. (2025) · neural information processing systems · doi:10.52202/085713-4899 · Cited by 148 · see citing works
Deep Structural Causal Models for Tractable Counterfactual InferencePreprint
Pawlowski, N., Castro, D. C., & Glocker, B. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2006.06485 · Cited by 96 · see citing works
Otter: A Multi-Modal Model with In-Context Instruction TuningPreprint
Li, B., Zhang, Y., Chen, L., et al. (2023) · arXiv (Cornell University) · doi:10.48550/arxiv.2305.03726 · Cited by 86 · see citing works
Wu, Z., Qiu, L., Ross, A., Akyürek, E., Chen, B., Wang, B., Kim, N., Andreas, J., & Kim, Y. (2024) · Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers) · doi:10.18653/v1/2024.naacl-long.102 · Cited by 58 · see citing works
Tomayto, Tomahto. Beyond Token-level Answer Equivalence for Question Answering Evaluation
Bulian, J., Buck, C., Gajewski, W., Börschinger, B., & Schuster, T. (2022) · Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing · doi:10.18653/v1/2022.emnlp-main.20 · Cited by 44 · see citing works
VIGC: Visual Instruction Generation and Correction
Wang, B., Wu, F., Han, X. H., Peng, J., Zhong, H., Zhang, P., Dong, X., Li, W., Li, W., Wang, J., et al. (2024) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v38i6.28338 · Cited by 42 · see citing works
Visually-Prompted Language Model for Fine-Grained Scene Graph Generation in an Open World
Yu, Q., Li, J., Wu, Y., Tang, S., Ji, W., & Zhuang, Y. (2023) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.01971 · Cited by 34 · see citing works
Self-citationFACTUAL: A Benchmark for Faithful and Consistent Textual Scene Graph Parsing
Li, Z., Chai, Y., Zhuo, T. Y., Qu, L., Haffari, G., Li, F., Ji, D., & Tran, Q. H. (2023) · Findings of the Association for Computational Linguistics: ACL (Findings of ACL) · doi:10.18653/v1/2023.findings-acl.398 · Cited by 26 · see citing works
Incorporating Bias-aware Margins into Contrastive Loss for Collaborative FilteringPreprint
Zhang, A., Ma, W., Wang, X., et al. (2022) · arXiv (Cornell University) · doi:10.48550/arxiv.2210.11054 · Cited by 17 · see citing works
COCO-Counterfactuals: Automatically Constructed Counterfactual Examples for Image-Text PairsPreprint
Le, T., Lal, V., & Howard, P. (2023) · arXiv (Cornell University) · doi:10.48550/arxiv.2309.14356 · Cited by 5 · see citing works
Prompting Large Language Models for Counterfactual Generation: An Empirical StudyPreprint
Li, Y., Xu, M., Miao, X., et al. (2023) · arXiv (Cornell University) · doi:10.48550/arxiv.2305.14791 · Cited by 2 · see citing works
Flagged first, then by citations
Cited by
Sorted by citations
Limitations of large language models in clinical problem-solving arising from inflexible reasoning
Kim, J., Podlasek, A., Shidara, K., Liu, F., Alaa, A., & Bernardo, D. (2025) · Scientific Reports · doi:10.1038/s41598-025-22940-0 · Cited by 78 · see citing works
Hallucination of Multimodal Large Language Models: A Survey
Bai, Z., Wang, P., Xiao, T., He, T., Han, Z., Zhang, Z., & Shou, M. Z. (2024) · arXiv · doi:10.48550/arxiv.2404.18930 · Cited by 38 · see citing works
“For mitigating hallucination, recent works make attempts on data, including introducing negative data [73], introducing counterfactual data [117], and reducing noise and errors in existing dataset [105, 120].”
InfluentialLi, G., Xu, J., Zhao, Y., & Peng, Y. (2025) · 2025 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr52734.2025.00850 · Cited by 12 · see citing works
- Authors
- Yu, Q., Li, J., Wei, L., Pang, L., Ye, W., Qin, B., Tang, S., Tian, Q., & Zhuang, Y.
- Affiliations
- Zhejiang University (CN)
- Huawei Technologies (China) (CN)
- Chinese Academy of Sciences (CN)
- ORCID
- Venue
- 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
- Publisher
- IEEE
- Published
- 2024-06-16
- Pages
- 12944-12953
- ISSN
- 1063-6919 (ISSN-L)
- OpenAlex
- W4402703048
- Cited by
- 59
- References
- 35
- Citations by year
- 2026: 31 · 2025: 22 · 2024: 6
- FWCI
- 12.22
- Type
- conference paper
- Language
- en
- Open access
- green
- arxiv.org
- Locations
- IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- Funders
- National Natural Science Foundation of China (U2336212)
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Multimodal Machine Learning Applications
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
- Physical Sciences › Computer Science › Artificial Intelligence › Topic Modeling
- Keywords
- hallucination mitigation, multimodal large language models, hallucination detection, cross-checking, object hallucination, spurious correlations, LLaVA
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
