Survey and analysis of hallucinations in large language models: attribution to prompting strategies or model behavior
Open accessThe DOI matches its Crossref record; 39 of the 40 references it cites were found in a registry or an index.
Abstract
Hallucination in Large Language Models (LLMs) refers to outputs that appear fluent and coherent but are factually incorrect, logically inconsistent, or entirely fabricated. As LLMs are increasingly deployed in education, healthcare, law, and scientific research, understanding and mitigating hallucinations has become critical. In this work, we present a comprehensive survey and empirical analysis of hallucination attribution in LLMs. Introducing a novel framework to determine whether a given hallucination stems from not optimize prompting or the model's intrinsic behavior. We evaluate state-of-the-art LLMs—including GPT-4, LLaMA 2, DeepSeek, and others—under various controlled prompting conditions, using established benchmarks (TruthfulQA, HallucinationEval) to judge factuality. Our attribution framework defines metrics for Prompt Sensitivity (PS) and Model Variability (MV) , which together quantify the contribution of prompts vs. model-internal factors to hallucinations. Through extensive experiments and comparative analyses, we identify distinct patterns in hallucination occurrence, severity, and mitigation across models. Notably, structured prompt strategies such as chain-of-thought (CoT) prompting significantly reduce hallucinations in prompt-sensitive scenarios, though intrinsic model limitations persist in some cases. These findings contribute to a deeper understanding of LLM reliability and provide insights for prompt engineers, model developers, and AI practitioners. We further propose best practices and future directions to reduce hallucinations in both prompt design and model development pipelines.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Figures







Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
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 cited source says what the claim attributes to it.
In the citing paper
“Such errors undermine scholarly trust and threaten the integrity of scientific communication [8, 25].”
From CiteAudit: You Cited It, But Did You Read It? A Benchmark for Verifying Scientific References in the LLM Era (2026) · cites this paper as [8]
The check wasn’t confident either way, so it gives no verdict — never disputed.
In the citing paper
“Third, and most critically, the field lacks an unprompted control condition. Every existing study prompts models to produce citations and then measures fabrication rates, but without establishing whether models spontaneously generate citations in the absence of an explicit request [11].”
From How LLMs Cite and Why It Matters: A Cross-Model Audit of Reference Fabrication in AI-Assisted Academic Writing and Methods to Detect Phantom Citations (2026) · cites this paper as [11]
Experimental — accuracy not yet measured
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Gelman, A., Carlin, J. B., Stern, H. S., & Rubin, D. B. (2003) · doi:10.1201/9780429258480 · Cited by 27,276 · cited 3×
“BHM represents hallucination rates hierarchically with model-specific and prompt-specific parameters drawn from higher-level distributions (Gelman et al., 2013): where H ij is hallucination rate for model i under prompt j , α i , β j represent model-specific and prompt-specific effects, and γ ij…”
Language Models are Few-Shot Learners
Brown, T. B., Mann, B., Ryder, N., Subbiah, M., Kaplan, J., Dhariwal, P., Neelakantan, A., Shyam, P., Sastry, G., Askell, A., et al. (2020) · arXiv · doi:10.48550/arxiv.2005.14165 · Cited by 23,372 · cited 3×
“Models such as GPT-3 (Brown et al., 2020), GPT-4 (OpenAI, 2023b), LLaMA 2 (Touvron et al., 2023), Claude (Anthropic, 2023), DeepSeek (DeepSeek AI, 2023), and others have demonstrated extraordinary capabilities in zero-shot and few-shot learning tasks.”
(1988) · Cited by 16,582
Kaugeranna, R., Kaugeranna, E., & Aion, (. S. 4. 6. (2023) · DROPS (Schloss Dagstuhl – Leibniz Center for Informatics) · doi:10.4230/lipics.giscience.2023.43 · Cited by 13,690
Statistical Decision Theory and Bayesian Analysis
Berger, J. O. (1985) · Springer Series in Statistics · doi:10.1007/978-1-4757-4286-2 · Cited by 7,722
Survey of Hallucination in Natural Language Generation
Ji, Z., Lee, N., Frieske, R., Yu, T., Su, D., Xu, Y., Ishii, E., Bang, Y. J., Madotto, A., & Fung, P. (2023) · ACM Computing Surveys · doi:10.1145/3571730 · Cited by 5,025 · cited 7×
“High-confidence hallucinations—those that appear fluent and plausible but are factually incorrect—are particularly dangerous and difficult to detect automatically (Kadavath et al., 2022; Ji et al., 2023).”
InfluentialChain-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. V., & Zhou, D. (2022) · Advances in Neural Information Processing Systems 35 · doi:10.52202/068431-1800 · Cited by 4,505
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 · cited 9×
“Chain-of-Thought (CoT) prompting (Wei et al., 2022) improves reasoning transparency and factual correctness by encouraging step-wise output generation.”
InfluentialTraining 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
Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks
Lewis, P., Perez, E., Piktus, A., Petroni, F., Karpukhin, V., Goyal, N., Küttler, H., Lewis, M., Yih, W.-T., Rocktäschel, T., et al. (2020) · arXiv · doi:10.48550/arxiv.2005.11401 · Cited by 3,101 · cited 6×
“To counter these issues, Retrieval-Augmented Generation (RAG) (Lewis et al., 2020), Grounded pretraining (Zhang et al., 2023), and contrastive decoding techniques (Li et al., 2022) have been explored.”
Influential
Research works printed but not linked: 3
- Kazemi, 2023, Cohs: a dataset for evaluating factual consistency of summaries, arXiv [preprint], 10.4850/arXiv:2305.00038 · doi:10.4850/arxiv:2305.00038
- Nakano, 2021, Webgpt: browser-assisted question-answering with human feedback, arXiv [preprint], 10.48550/arXiv:2112.09332 · doi:10.48550/arxiv:2112.09332
- Zhang, 2023, Grounded language model training reduces hallucination, arXiv [preprint], 10.48550/arXiv:.303.08239 · doi:10.48550/arxiv:.303.08239
Cited by
Data Quality in the Age of AI: A Review of Governance, Ethics, and the FAIR Principles
Guillen-Aguinaga, M., Aguinaga-Ontoso, E., Guillen-Aguinaga, L., Guillen-Grima, F., & Aguinaga-Ontoso, I. (2025) · Data · doi:10.3390/data10120201 · Cited by 37
Next-generation probiotics: an outlook into current applications and future developments
Kern, L., Tofield, A., Frame, J., & Elinav, E. (2026) · Nature Reviews Microbiology · doi:10.1038/s41579-026-01311-0 · Cited by 31
“Systematic evaluation of state-of-the-art models revealed that hallucinations arise from both prompt-related ambiguity and intrinsic model behaviour, even under controlled prompting strategies and evaluation benchmarks such as TruthfulQA and HallucinationEval (now known as HalluClean) 180–182 .”
Earp, B. D., Frith, L., Mann, S. P., Sawai, T., Shahvisi, A., & Wangmo, T. (2026) · Journal of Medical Ethics · doi:10.1136/jme-2026-112179 · Cited by 16
Hodzic, S., Stević, A., & Matthes, J. (2026) · Technology in Society · doi:10.1016/j.techsoc.2026.103331 · Cited by 12
Hill, C., Dahil, A., Simpson, G., Hardisty, D., Keast, J. M., Pinn, C. K., & Dambha‐Miller, H. (2026) · PLOS Digital Health · doi:10.1371/journal.pdig.0001189 · Cited by 16
Metadata
- Authors
- Anh-Hoang, D., Tran, V., & Nguyen, L.-M.
- Affiliations
- Japan Advanced Institute of Science and Technology (JP)
- Corresponding author
- Anh-Hoang, Dang
- Venue
- Frontiers in Artificial Intelligence
- Publisher
- Frontiers Media SA
- Published
- 2025-09-30
- Volume
- 8
- Pages
- 1622292-1622292
- ISSN
- 2624-8212 (ISSN-L)
- PMC
- PMC12518350
- OpenAlex
- W4414677042
- Semantic Scholar
- a79eb7fddd1b620b12f41243b8c139f01f7b4e4e
- Cited by
- 146
- References
- 27
- Citations by year
- 2026: 140 · 2025: 5
- FWCI
- 13.91
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- Locations
- Frontiers in Artificial Intelligence (publishedVersion, cc-by)
- Europe PMC (PubMed Central) (publishedVersion, cc-by)
- PubMed (publishedVersion)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- Funders
- Japan Society for the Promotion of Science (23K16954)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Topic Modeling
- Physical Sciences › Computer Science › Artificial Intelligence › Natural Language Processing Techniques
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Multimodal Machine Learning Applications
- Keywords
- hallucinations, large language models, prompting strategies, prompt sensitivity, model variability, chain-of-thought prompting, factual accuracy, hallucination mitigation, prompt engineering, TruthfulQA
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
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: Survey and analysis of hallucinations in large language models: attribution to prompting strategies or model behavior. https://app.cochat.ai/papers/10.3389/frai.2025.1622292
- The paper
- Anh-Hoang, D., Tran, V., & Nguyen, L.-M. (2025). Survey and analysis of hallucinations in large language models: attribution to prompting strategies or model behavior. Frontiers in Artificial Intelligence. https://doi.org/10.3389/frai.2025.1622292
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.