Survey of Hallucination in Natural Language Generation
Open accessThe DOI matches its Crossref record; 355 of the 360 references it cites were found in a registry or an index, and 4 are not checked yet; 224 of the 256 sources its claims cite were checked for support.
Abstract
Natural Language Generation (NLG) has improved exponentially in recent years thanks to the development of sequence-to-sequence deep learning technologies such as Transformer-based language models. This advancement has led to more fluent and coherent NLG, leading to improved development in downstream tasks such as abstractive summarization, dialogue generation, and data-to-text generation. However, it is also apparent that deep learning based generation is prone to hallucinate unintended text, which degrades the system performance and fails to meet user expectations in many real-world scenarios. To address this issue, many studies have been presented in measuring and mitigating hallucinated texts, but these have never been reviewed in a comprehensive manner before. In this survey, we thus provide a broad overview of the research progress and challenges in the hallucination problem in NLG. The survey is organized into two parts: (1) a general overview of metrics, mitigation methods, and future directions, and (2) an overview of task-specific research progress on hallucinations in the following downstream tasks, namely abstractive summarization, dialogue generation, generative question answering, data-to-text generation, and machine translation. This survey serves to facilitate collaborative efforts among researchers in tackling the challenge of hallucinated texts in NLG.
Verified 2026-11-19 · 4 references not checked yet
Overview
CoChat Summary
TL;DR
This survey provides a broad overview of hallucination in natural language generation, covering definitions, contributors, metrics, mitigation methods, and task-specific progress across summarization, dialogue, GQA, data-to-text, machine translation, vision-language generation, and LLMs. It distinguishes intrinsic from extrinsic hallucination and argues that hallucination must be treated differently across tasks and sources.
Key findings
- The survey categorizes hallucination into intrinsic hallucination, where generated output contradicts the source, and extrinsic hallucination, where output cannot be verified from the source; extrinsic hallucination is not always erroneous because it may come from factually correct external information.
- It reports prevalence examples from cited work: 25% of summaries generated by state-of-the-art abstractive summarization systems have hallucinated content; 62% of first sentences in WIKIBIO have additional information not stated in the corresponding infobox; and 74% of samples in the QMSum dataset consist of inconsistent facts.
- It reviews hallucination metrics as statistical, model-based (IE-based, QA-based, NLI-based, faithfulness classification, LM-based), and human evaluation, and mitigation methods as data-related and modeling/inference methods including architecture changes, planning/sketching, RL, multi-task learning, controllable generation, and post-processing.
- For LLMs, it notes that RLHF reduced the hallucination rate from 41% to 21% in closed-domain tasks for InstructGPT, and discusses reference-dependent and reference-free hallucination metrics plus mitigation via data refinement, RAG, safety fine-tuning, model editing, decoding methods, chain-of-thought variants, post-processing, and ensembling.
Methods
- Literature survey design: the paper reviews existing research on hallucination in NLG rather than conducting new experiments or primary data collection.
- Coverage: general NLG hallucination topics (definitions, contributors, metrics, mitigation, future directions) plus task-specific sections on abstractive summarization, dialogue generation, generative question answering, data-to-text generation, neural machine translation, vision-language generation, and large language models.
- Analysis: synthesizes and compares definitions, metrics, mitigation approaches, and future directions across tasks, with examples in tables and figures; the LLM section was updated in Jan 2024.
Limitations
- The paper states that GQA has no standard definition or categorization of hallucination and no automatic metric specifically for hallucination in GQA; it also notes that open-domain dialogue lacks a standard hallucination metric and that VL hallucination research is still in its early stage.
- The paper states that extrinsic hallucination has been somewhat overlooked and is more challenging to reduce, and that most existing mitigation methods do not distinguish between intrinsic and extrinsic hallucination.
- Evident from the text: as a survey, it does not provide new empirical validation of metrics or mitigation methods, and its LLM coverage is explicitly updated only through Jan 2024, so later work is not included.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Figures
Verification
Claims the paper makes
124 Valid · 26 Not valid · 19 Not checked · 135 Own results and interpretations
At least one cited source doesn’t support this claim or points to a different paper, and none confirmed it — see the verdicts below.
In the paper
“Furthermore, perturbation, where the insertions of most common tokens are placed at the beginning of the sentence, seems to be the most successful in hallucination mitigation.”
What the sources say
- Not checked — no full text of the sourcecites [133] [133] Katherine Lee, Orhan Firat, Ashish Agarwal, Clara Fannjiang, and David Sussillo. 2019. Hallucinations in Neural Machine Translation. ICLR (2019). -
- Disputedcites [215] [215] Vikas Raunak, Arul Menezes, and Marcin Junczys-Dowmunt. 2021. The Curious Case of Hallucinations in Neural Machine Translation. (4 2021), 1172-1183. https://arxiv.org/abs/2104.06683v1 - sourceExperimental — accuracy not yet measured
In the passage describing the perturbation procedure (page unavailable), the paper says that inserting common tokens at the beginning is used to generate hallucinations, not mitigate them. Its prevention section instead discusses approaches such as robust learning and corpus-level filtering. (LLM, conf=0.92)
Checked by a reasoning model — may be wrong.
hedgedstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
At least one cited source doesn’t support this claim or points to a different paper, and none confirmed it — see the verdicts below.
In the paper
“While there are works on the visual extension of faithfulness-oriented RLHF [303], these essentially involve labeling more data, which is expensive and difficult to scale.”
What the sources say
- Disputedcites [303] [303] Tianyu Yu, Yuan Yao, Haoye Zhang, Taiwen He, Yifeng Han, Ganqu Cui, Jinyi Hu, Zhiyuan Liu, Hai-Tao Zheng, Maosong Sun, and Tat-Seng Chua. 2023. RLHF-V: Towards Trustworthy MLLMs via Behavior Alignment from Fine-grained Correctional Human Feedback. CoRR abs/2312.00849 (2023). https://doi.org/10.48550/ARXIV.2312.00849 arXiv:2312.00849 - sourceExperimental — accuracy not yet measured
RLHF-V does use human annotation for visual-model alignment, but it makes the contrary point about needing more labeled data. The Fine-grained Correctional Human Preference Collection section says its corrections improve data efficiency, and Section 4.3 reports comparable hallucination rates using only 200 preference examples versus an order of magnitude more ranking-labeled examples. Our copy does not establish a separate claim about annotation cost or scalability. (LLM, conf=0.88)
Checked by a reasoning model — may be wrong.
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Disputed
Every source the paper cites for this claim was checked and supports it.
In the paper
“Nevertheless, as illustrated by Falke et al. [66], off-the-shelf NLI models tend to transfer poorly to the abstractive summarization task.”
What the sources say
- Supportedcites [66] [66] Tobias Falke, Leonardo FR Ribeiro, Prasetya Ajie Utama, Ido Dagan, and Iryna Gurevych. 2019. Ranking Generated Summaries by Correctness: An Interesting but Challenging Application for Natural Language Inference. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics . 2214-2220. - source
The Introduction states that models trained on NLI datasets transfer poorly to the paper’s downstream task of ranking generated summaries by correctness. The Conclusions likewise say that out-of-the-box NLI models do not perform well on that task. (LLM, conf=0.94)
- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Such wrong association results in generation with facts mixed up between two similar entities [60, 233].”
What the sources say
- Partly supportedcites [60] [60] Nouha Dziri, Andrea Madotto, Osmar Zaiane, and Avishek Joey Bose. 2021. Neural Path Hunter: Reducing Hallucination in Dialogue Systems via Path Grounding. EMNLP (2021). - sourceExperimental — accuracy not yet measured
The GPT2-KG passage describes an incorrect association between entities, and G Error Analysis gives an example of a generated response assigning a book to the wrong author. These support the claim that facts can be mixed up, but the paper does not say the entities involved are similar. The supplied passages do not provide page numbers. (LLM, conf=0.83)
- Supportedcites [233] [233] Kurt Shuster, Spencer Poff, Moya Chen, Douwe Kiela, and Jason Weston. 2021. Retrieval Augmentation Reduces Hallucination in Conversation. EMNLP (2021). - source
The Introduction states that language models often mix up facts between two similar entities when generating factually incorrect statements. This directly supports the cited description of the resulting error. (LLM, conf=0.92)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“As a result, the target reference may contain information that cannot be supported by the source [198, 268].”
What the sources say
- Supportedcites [198] [198] Ankur Parikh, Xuezhi Wang, Sebastian Gehrmann, Manaal Faruqui, Bhuwan Dhingra, Diyi Yang, and Dipanjan Das. 2020. ToTTo: A Controlled Table-To-Text Generation Dataset. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP) . 1173-1186. - source
The Dataset section says that reference targets selected from web sentences may contain information that cannot be inferred from the source. The Data Annotation Process section also describes deleting phrases unsupported by table cells. (LLM, conf=0.93)
- Supportedcites [268] [268] Hongmin Wang. 2019. Revisiting Challenges in Data-to-Text Generation with Fact Grounding. In Proceedings of the 12th International Conference on Natural Language Generation . 311-322. - source
The Introduction reports that about 40% of human-written game-summary content cannot be directly mapped to the input boxscore records and draws on information beyond those inputs. Section 2.1 likewise describes reference content not licensed by the boxscore table. This supports the claim that a target reference may contain information the source does not support. (LLM, conf=0.93)
hedgedstates a possible cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
117 more claims · 74 with their cited sources checked
Create a free account to see all 304 claims and to verify the 2 references still open.
- Borderline: one of the tests for a substantial claim is unclear. 1
- It may state two separate claims as one. 90
- Probably not a substantial claim. 142
- The quote only partly states this claim. 1
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 source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Hallucination can lead to serious consequences such as incorrect or misleading code, security risks, and ethical issues [186].”
From A survey on large language model based autonomous agents (2024) · cites this paper as [186]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“citation fabrication constitutes a form of extrinsic hallucination in the absence of retrieval augmentation, since the model generates references not grounded in any provided source material [4].”
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 [4]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Recent taxonomic work distinguishes at least seven cases [15,18].”
From Hallucination as epistemic pollution: epistemic injustice and credibility in futures-oriented science education (2026) · cites this paper as [18]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“LLMs struggle to generate accurate and up-to-date responses on current topics, due to outdated knowledge, lack of domain-specific expertise, or hallucination13–17.”
From Leveraging long context in retrieval augmented language models for medical question answering (2025) · cites this paper as [16]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Fine-tuning can enrich LLMs with domain knowledge, enhance their specific ability, improve the fairness and reliability of the outputs, and prevent certain damage caused by hallucination [26].”
From FederatedScope-LLM: A Comprehensive Package for Fine-tuning Large Language Models in Federated Learning (2024) · cites this paper as [26]
Experimental — accuracy not yet measured
41 more claims about this paper
Create a free account to see every claim other papers make about this one.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Extrinsic Object Hallucination
“CHAIR (Caption Hallucination Assessment with Image Relevance) metric, which calculates what proportion of object words generated are actually in the image according to the ground truth captions.”
§ Extrinsic Object Hallucination
“CHAIR 𝑖 measures per-instance object hallucination, i.e. what fraction of object instances in each generated caption are hallucinated.”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ 7.1 Hallucination Definition in Abstractive Summarization
“For instance, in Table 1, given the input article shown in the caption, an example of intrinsic hallucination is ' The Ebola vaccine was rejected by the FDA in 2019, '”
§ 2.1 Categorization
“in the abstractive summarization task from Table 1, the generated summary ' The first Ebola vaccine was approved in 2021 ' contradicts the source content”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Survey of Hallucination in Natural Language Generation
“The survey is organized into two parts: (1) a general overview of metrics, mitigation methods, and future directions; (2) an overview of task-specific research progress on hallucinations in the following downstream tasks, namely abstractive summarization, dialogue generation, generative question answering, data-to-text generation, machine translation, and visual-language generation; and (3) hallucinations in large language models (LLMs)”
§ 1 INTRODUCTION
“The third part discusses this phenomenon in LLMs in Section 13.”
Citations
Cites
Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, L., & Polosukhin, I. (2017) · arXiv · doi:10.48550/arxiv.1706.03762 · Cited by 120,404
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
Devlin, J., Chang, M.-W., Lee, K., & Toutanova, K. (2019) · Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers) · doi:10.18653/v1/n19-1423 · Cited by 54,325 · cited 3×
“Recently, the rapid development of NLG has captured the imagination of many thanks to the advances in deep learning technologies, especially Transformer [189]-based models like BERT [29], BART [100], GPT-2 [149], and GPT-3 [16].”
BackgroundAn Image is Worth 16x16 Words: Transformers for Image Recognition at Scale
Dosovitskiy, A., Beyer, L., Kolesnikov, A., Weissenborn, D., Zhai, X., Unterthiner, T., Dehghani, M., Minderer, M., Heigold, G., Gelly, S., et al. (2020) · arXiv · doi:10.48550/arxiv.2010.11929 · Cited by 47,472
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
Sequence to Sequence Learning with Neural Networks
Sutskever, I., Vinyals, O., & Le, Q. V. (2014) · arXiv · doi:10.48550/arxiv.1409.3215 · Cited by 17,766
Advances in Neural Information Processing Systems 29
Teymur, O., Zygalakis, K., & Calderhead, B. (2016) · Cited by 13,887
Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer
Raffel, C., Shazeer, N., Roberts, A., Lee, K., Narang, S., Matena, M., Zhou, Y., Li, W., & Liu, P. J. (2019) · arXiv · doi:10.48550/arxiv.1910.10683 · Cited by 10,929 · cited 1×
“This work investigates the performances of GPT-3 [13], GPT-Neo/J [141], GPT-2 [105] and a T5-based model [106].”
BackgroundAdvances in Neural Information Processing Systems 14: Proceedings of the 2001 Conference
(2002) · The MIT Press eBooks · doi:10.7551/mitpress/1120.001.0001 · Cited by 8,890
Lewis, M., Liu, Y., Goyal, N., Ghazvininejad, M., Mohamed, A., Levy, O., Stoyanov, V., & Zettlemoyer, L. (2020) · Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics · doi:10.18653/v1/2020.acl-main.703 · Cited by 8,821 · cited 1×
“Recently, the development of NLG has conquered the imagination of many thanks to the advances in deep learning technologies, especially the Transformer [159]-based models like BERT [24], BART [83], GPT-2 [127], and GPT3 [13].”
BackgroundLewis, M., Liu, Y., Goyal, N., Ghazvininejad, M., Mohamed, A., Levy, O., Stoyanov, V., & Zettlemoyer, L. (2019) · arXiv · doi:10.48550/arxiv.1910.13461 · Cited by 8,816 · cited 1×
“Recently, the development of NLG has conquered the imagination of many thanks to the advances in deep learning technologies, especially the Transformer [159]-based models like BERT [24], BART [83], GPT-2 [127], and GPT3 [13].”
Background
Research works printed but not linked: 16
- [4] Joshua Albrecht and Rebecca Hwa. 2007. A Re-examination of Machine Learning Approaches for Sentence-Level MT Evaluation. In Proceedings of the 45th Annual Meeting of the Association of Computational Linguistics . Association for Computational Linguistics, Prague, Czech Republic, 880-887. https://aclanthology.org/P07-1111 -
- [16] Andrzej Białecki, Robert Muir, Grant Ingersoll, and Lucid Imagination. 2012. Apache lucene 4. In SIGIR 2012 workshop on open source information retrieval . 17. -
- [53] Yilun Du, Shuang Li, Antonio Torralba, Joshua B Tenenbaum, and Igor Mordatch. 2023. Improving Factuality and Reasoning in Language Models through Multiagent Debate. In Forty-first International Conference on Machine Learning . -
- [121] Philipp Koehn and Rebecca Knowles. 2017. Six Challenges for Neural Machine Translation. In First Workshop on Neural Machine Translation . Association for Computational Linguistics, 28-39. -
- [145] Junnan Li, Dongxu Li, Caiming Xiong, and Steven Hoi. 2022. BLIP: Bootstrapping Language-Image Pre-training for Unified Vision-Language Understanding and Generation. In ICML . -
- [149] Tian Li, Ahmad Beirami, Maziar Sanjabi, and Virginia Smith. 2020. Tilted Empirical Risk Minimization. In International Conference on Learning Representations . -
- [164] Jiasen Lu, Jianwei Yang, Dhruv Batra, and Devi Parikh. 2018. Neural Baby Talk. In CVPR . -
- [165] Yukun Ma, Khanh Linh Nguyen, Frank Z Xing, and Erik Cambria. 2020. A Survey on Empathetic Dialogue Systems. Information Fusion 64 (2020), 50-70. -
- [175] Marianna J. Martindale, Marine Carpuat, Kevin Duh, and Paul McNamee. 2019. Identifying Fluently Inadequate Output in Neural and Statistical Machine Translation. In MTSummit . -
- [179] Kathleen McKeown. 1992. Text Generation . Cambridge University Press. -
- [209] Alec Radford, Jeffrey Wu, Rewon Child, David Luan, Dario Amodei, and Ilya Sutskever. 2019. Language Models are Unsupervised Multitask Learners. OpenAI Blog 1, 8 (2019), 9. -
- [238] Hui Su, Xiaoyu Shen, Sanqiang Zhao, Zhou Xiao, Pengwei Hu, Cheng Niu, and Jie Zhou. 2020. Diversifying Dialogue Generation with Non-Conversational Text. In 58th Annual Meeting of the Association for Computational Linguistics . ACL, 7087-7097. -
- [256] Zhaopeng Tu, Yang Liu, Lifeng Shang, Xiaohua Liu, and Hang Li. 2017. Neural Machine Translation with Reconstruction. In Thirty-First AAAI Conference on Artificial Intelligence . -
- [261] Neeraj Varshney, Wenlin Yao, Hongming Zhang, Jianshu Chen, and Dong Yu. 2023. A Stitch in Time Saves Nine: Detecting and Mitigating Hallucinations of LLMs by Validating Low-Confidence Generation. CoRR abs/2307.03987 (2023). https://doi.org/10.48550/arXiv.2307.03987 arXiv:2307.03987 - · arXiv · doi:10.48550/arxiv.2307.03987 · Paper page
- [285] Zeqiu Wu, Michel Galley, Chris Brockett, Yizhe Zhang, Xiang Gao, Chris Quirk, Rik Koncel-Kedziorski, Jianfeng Gao, Hannaneh Hajishirzi, Mari Ostendorf, et al. 2021. A Controllable Model of Grounded Response Generation. In Proceedings of the AAAI Conference on Artificial Intelligence , Vol. 35. 14085-14093. -
- [291] Xinnuo Xu, Ondřej Dušek, Verena Rieser, and Ioannis Konstas. 2021. AGGGEN: Ordering and Aggregating while Generating. roceedings of the 59th Annual Meeting of the Association for Computational Linguistics (ACL2021) (2021). -
Cited by
Items 4,401–4,500 of 4,844
BAYRAKLI, V. K., Demircioğlu, A., & Ertek, Z. (2026) · Advances in computational intelligence and robotics book series · doi:10.4018/979-8-2600-1232-1.ch002
Navarro-Bringas, E. (2026) · Journal of Information Technology in Construction · doi:10.36680/j.itcon.2026.042
Esfahlani, S. S., Prothero, L., Abdulmohdi, N., Nantheswaran, N., Turvey, J., & Jacobs, C. (2026) · JMIR Formative Research · doi:10.2196/100237
“…is particularly relevant in adult autism assessment, where fixed-response questionnaires may not adequately capture the complexity and heterogeneity of adult presentations, especially in individuals who engage in masking behaviors or present with co-occurring psychiatric conditions [4,13,15,24].”
Horr, A., Walther, I. L., Bonietzki, A., Piesch, T.-C., Peters, S., Sarikaya, M. M., Seehafer, S., Jansen, O., Langguth, P., & Aludin, S. (2026) · Insights into Imaging · doi:10.1186/s13244-026-02391-5
Intelligent Industrial Scene Interpretation via Synergy of Deep Learning and Knowledge Embedding
Hu, F., & Hu, Z. (2026) · IEEE Access · doi:10.1109/access.2026.3728105
“…at the reasoning layer, as the generative reasoning nature of large models is inherently probabilistic and cannot provide deterministic guarantees [12]; and the absence of structured attribution at the output layer, as large models struggle to produce legally formatted damage-cause statistical…”
Wang, Y., Li, M., Liu, S., & Xue, L. (2026) · Expert Systems with Applications · doi:10.1016/j.eswa.2026.134222
Attunement Costs: How Interaction Locks Users into Specific LLMsPreprint
Hanschur, L., Obermeier, D., & Henkel, J. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.7354642
Huynh, C. N., & Nguyen, L. T. (2026) · International Conference on Education and Training Technologies (ICETT) · doi:10.1109/icett70124.2026.11665998
“3) Situational Failure of Habits: While habits are often stable traits, real-time interaction with a highly persuasive and fluent GenAI may neutralize these safeguards [3].”
InfluentialMawardi, V. C., Purwarianti, A., Trilaksono, B. R., Chayadi, J., & Tiatri, S. (2026) · doi:10.1109/icoict71118.2026.11664855
“This limits the ability of LLMs to reason based on source text, resulting in models tending to produce generic or hallucinatory answers [9], [10], [11].”
Boldi, A., Gabbatore, I., & Bosco, F. M. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.7366891
Wissensbasierte Qualitätssicherung durch KI
Weisbrod, N., Filling, J. M., Bolz, H., & Metternich, J. (2026) · Zeitschrift für wirtschaftlichen Fabrikbetrieb · doi:10.1515/zwf-2026-1097
Risk-aware multiple-choice question answering with conformal P-values
Fu, W., Ye, Y., Yang, G., Wei, Y., & Zhang, L. (2026) · Scientific Reports · doi:10.1038/s41598-026-66677-w
A Generative AI-Driven IoT System Design Validation Framework for Multiple Sensors
Kong, D., Funabiki, N., Kotama, I. N. D., Zhu, Z., & Kao, W.-C. (2026) · International Conference on Computer Communication and the Internet (ICCCI) · doi:10.1109/iccci70321.2026.11666561
“When dealing with strict hardware constraints and multi-source engineering documents, conventional RAG architectures frequently struggle with reasoning logic and are highly susceptible to generative hallucinations [6].”
InfluentialMukkawar, S., Pachpute, S., Paul, O., Paul, K., & Deshmukh, P. (2026) · International Conference on Intelligent and Sustainable Electronics & Computing Technologies (INSECT) · doi:10.1109/insect68872.2026.11663745 · Cited by 1
“Ji et al. [1] categorize hallucinations into two types: intrinsic hallucinations (contradicting the source) and extrinsic hallucinations (adding unverifiable information).”
Traceability and Source Authority Guardrails for Agentic RAG-Based Financial Claim Verification
Thamrin, D. S., Lubis, M., Fansyah, E. A., Fauzan, R. A., & Asyam, M. R. D. (2026) · International Conference on Electronics Representation and Algorithm (ICERA) · doi:10.1109/icera72709.2026.11666732
“Recent studies have proposed methods to improve retrieval consistency [26] and address retrieval security risks [27], while hallucination research shows that fluent responses are not always supported by evidence [4], [5].”
Kim, B., & Yu, Y. (2026) · Nuclear Engineering and Technology · doi:10.1016/j.net.2026.104687
Mahmud, Z. F. N., Bryan, Dadas, I. A., & Kristiana, I. (2026) · International Conference on Electronics Representation and Algorithm (ICERA) · doi:10.1109/icera72709.2026.11666711
ADAPTIVE TURKISH FEDERATED RAG ARCHITECTURE FOR LOW-RESOURCE SYSTEMS
Toy, M. T., & Süzen, A. A. (2026) · International Journal of 3D Printing Technologies and Digital Industry · doi:10.46519/ij3dptdi.1895770
EG-MAP: Evidence-Grounded Multi-Agent Smart Contract Auditing
Song, Z., Yue, L., 陶泽家, Han, Q., Zeng, W., Xu, Z., Pal, S., & Dong, C. (2026) · IEEE International Conference on Software Services Engineering (SSE) · doi:10.1109/sse72781.2026.00018
“By contrast, unguided LLM-based auditing may hallu-cinate plausible but unsupported attacks [9], [10].”
Xu, J., & Zhang, Z. (2026) · International Conference on Distance Education and Learning (ICDEL) · doi:10.1109/icdel70250.2026.11663645
“This allows educators to rapidly assemble rich, varied, and contextually relevant teaching resources, freeing their intellectual and temporal capital for higher-order tasks: designing compelling learning sequences, facilitating nuanced in-class discussions, and providing targeted mentorship [7].”
Pérez-Navarro, J., & Navarro-Fernández, S. (2026) · Anales de Edificación · doi:10.20868/ade.2026.5724
Silva, A., & Lino, M. (2026) · Nursing Inquiry · doi:10.1111/nin.70162
“As Thorne and others have observed within the nursing inquiry tradition, the framing of a problem shapes which solutions appear available, which makes the governance of problem framing in AI outputs a nursing‐specific concern (Jermutus et al. 2022; Ji et al. 2023).”
Beyond One-Way Pipelines: Bidirectional Cooperation for Self-explaining RationalizationPreprint
Deng, Z., Zheng, L., Gao, W., Yuan, W., Liu, D., & Liu, W. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.7380313
Park, Y., Kim, C., Kim, G., et al. (2026) · Computers and Electronics in Agriculture · doi:10.1016/j.compag.2026.112347
Arfiansyah, M. A., Kamal, M., & Fiddina, Q. A. (2026) · International Conference on Electronics Representation and Algorithm (ICERA) · doi:10.1109/icera72709.2026.11666677
“Despite their effectiveness, LLMs suffer from hallucination and knowledge cut-off limitations, restricting their reliability in dynamic and domain-specific tasks [4], [12].”
Tani, T., Joseph, M. H., Metso, L., & Tynninen, L. (2026) · Computer Applications in Engineering Education · doi:10.1002/cae.70263
“A further subset cited “just check the manual” responses (8) or contradictory answers across attempts (4), echoing failure modes documented in the RAG literature [12, 46, 47, 71].”
InfluentialXia, D., Xu, P., Shi, J., Zhang, F., Tan, J., Wan, J., & Li, B. (2026) · Journal of Systems Architecture · doi:10.1016/j.sysarc.2026.103975
Nguyen, T.-N.-L., Hoang, A.-D., Tran, M.-C., Vu, M.-T., Nguyen, X.-B., Le, H.-Q., Vuong, T.-H.-Y., Phan, X.-H., & Tran, M.-V. (2026) · IEEE Access · doi:10.1109/access.2026.3729218
“Moreover, LLMs are trained on a fixed dataset, and their inability to acquire new knowledge leads to outdated, incorrect information about recent developments [37].”
Practical Research on Building Professional Knowledge Q&A Robot Based on RAG
Zhou, H., Zhao, L., Tan, L., & Li, Y. (2026) · International Conference on Computer Information and Big Data Applications (CIBDA) · doi:10.1145/3813822.3814122
Ramani, A., & Venugopalan, M. (2026) · Discover Artificial Intelligence · doi:10.1007/s44163-026-02106-1
“A more comprehensive review of Hallucinations for various tasks in NLG has been done in [51], covering downstream tasks such as abstractive summarization, dialog generation, machine translation, and question-answering.”
Xie, C., Ren, S., Ouyang, H., Wang, C., Bonilla, D., & Liao, X. (2026) · Sustainability · doi:10.3390/su18178932
Multi-Path Causal Optimization for claim verification through controlling confounding
Guo, H., Di, S., Meo, P. D., Chen, Z., & Zhu, J. (2026) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2026.116117
AI-agents in de onderzoekscyclus: een praktische handleiding voor het versnellen van scoping reviews
Witte, L. P. W., Ribbert, L., Frijns, F., & van Toorenburg, E. (2026) · Tijdschrift voor Urologie · doi:10.1007/s13629-026-00507-6
Role-Aware Agentic Retrieval-Augmented Generation for Urban Rail Transit Fault Management
Zhang, W., Fang, H., Bu, J., Zhong, Q., & Yu, Y. (2026) · AI · doi:10.3390/ai7090339
“However, directly applying LLMs to URT operations remains challenging because their open-ended generation may produce unsupported recommendations that deviate from established operational procedures [5,6].”
Lim, J., & Eom, S.-J. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.7395408
Generative large language models in medicine: a scoping review of recent methodological advances
LI, F., Li, J., Cao, W., Li, H., Li, Y., Sun, Z., Chen, D., Duan, T., Li, P., Tizhoosh, H., et al. (2026) · npj Health Systems · doi:10.1038/s44401-026-00137-5 · Cited by 1
“In such settings, models can generate plausible-sounding but unsupported statements, fabricate diagnoses, or con fl ate temporal events, creating substantial risks for downstream decision-making 178 .”
Sun, Y., Liu, D., Zhang, J., et al. (2026) · Computers · doi:10.3390/computers15090576
GenAI and Corpus-Driven Indonesian L2 Writing
Qin, W. (2026) · International Conference on Innovation Management and Information Systems (ICIMIS) · doi:10.1145/3821966.3822127
Aljohani, M., Кеčo, D., Jawad, M. S., & Alzahrani, T. (2026) · International Journal of E-Services and Mobile Applications · doi:10.4018/ijesma.420859 · Cited by 1
Nasrin, F., Uddın, M. M., & Luo, F. (2026) · Education Sciences · doi:10.3390/educsci16091426
Aftiss, A., Lamsiyah, S., Schommer, C., & Alaoui, S. O. E. (2026) · IEEE Access · doi:10.1109/access.2026.3729844
“Hallucination, considered as the generation of content not grounded in the source document, represents a critical risk in LLM-based summarization, particularly in the biomedical domain where factual accuracy is essential [87].”
Predictive Analysis of User Experience in Conversational Design Applications
Venugopal, K. (2026) · Advances in computational intelligence and robotics book series · doi:10.4018/979-8-2600-0512-5.ch004
Ren, J., Yang, Z., Liu, W., & Ng, M. T. A. (2026) · Briefings in Bioinformatics · doi:10.1093/bib/bbag456
“LLMs are also prone to fluent but weakly supported explanations when not constrained by structured evidence [ 23].”
Kumar, G., & Sharma, S. (2026) · Advances in computational intelligence and robotics book series · doi:10.4018/979-8-3373-9200-4.ch001
Integrating Large Language Models in Socially Assistive Robots for Personalized Eldercare
Villa, L., Hervás, R., Cubero, G., Cabañero, L., & Carneros-Prado, D. (2026) · International Journal of Social Robotics · doi:10.1007/s12369-026-01436-5
“…evaluate the quality of conversations generated by these robotic agents to ensure they are appropriate and well-perceived by users, particularly older adults, as LLMs still present significant limitations in their ability to gener - ate coherent and factual responses in open-domain dialogue [19].”
Sawad, A. B., & Binsawad, M. H. (2026) · Applied Sciences · doi:10.3390/app16178779
Accountable Autonomy: A Governance Framework for Agentic AI in Telecommunication Networks
Uriol, J., O'Brien, E., Hernández, I., Viola, R., Iradier, E., & Montalbán, J. (2026) · AI · doi:10.3390/ai7090347 · Cited by 2
“Prior work has shown that LLM-driven agents may suffer from hallucination, schema drift, and inconsistent decision patterns [52,53], which can translate into incorrect or unsafe operational choices.”
Jôrgensen, B. N., & Ma, Z. G. (2026) · Automation · doi:10.3390/automation7050138
Turaev, S., John, M., Awad, M., Zaki, N., & Shuaib, K. (2026) · IEEE Access · doi:10.1109/access.2026.3730379 · Cited by 2
“The use of language models in a governance instrument cannot be naive, because these models hallucinate , producing confident outputs that are factually unsupported [56], and a hallucinated competency record could misdirect curricular change.”
Prakash, D., Kulkarni, P., Rangarajan, V., & Kumar, M. (2026) · Indian Journal of Nuclear Medicine · doi:10.25259/ijnm_104_2026
Wang, Z.-Q., Zhou, Y., Song, W., Yao, X.-J., Chen, K.-F., Liu, J., & Xue, D.-F. (2026) · Advanced Functional Materials · doi:10.1002/adfm.78176
Guo, C., Zheng, J., Su, P., Tian, Z., Tang, J., Li, S., & Wang, T. (2026) · Information Processing & Management · doi:10.1016/j.ipm.2026.105112
CRAFT: Corpus Relatedness Analysis using Fourier Transforms
Chen, K., & Koudas, N. (2026) · Proceedings of the VLDB Endowment · doi:10.14778/3828612.3828622
Shinohara, T. (2026) · Street Art & Urban Creativity · doi:10.62161/sauc.v12.6387
“An agent can retrieve records, summarize evidence, call predictive tools, and propose plans through a natural-language interface; retrieval-augmented generation can ground such responses in external records, but hallucination remains a documented failure mode (Lewis et al., 2020; Ji et al., 2023).”
Beyond the autonomy illusion: hybrid intelligent systems and the architecture of AI accountability
Medsker, L., & Virmani, S. (2026) · AI and Ethics · doi:10.1007/s43681-026-01327-5
“…language models generate fabricated legal citations with high confidence [ 6], produce fluent but factually incorrect clinical recommendations [ 8], and exhibit bias and opacity that resist correction because the relevant properties are distributed across billions of weights rather than…”
Growth Decision Agent Based on Graph Retrieval Enhancement
Yuan, L. (2026) · International Conference on Cyber Security, Artificial Intelligence and Digital Economy (CSAIDE) · doi:10.1145/3819836.3819978
“…systematically explored the illusion problem in generative models, showing that the lack of reliable external knowledge constraints is a key factor [5]; Lavri-novics et al. explored the potential to alleviate illusion from the perspective of the association between knowledge graph and language…”
A survey on synergizing large language models and knowledge graphs for reasoning
Wang, X., Bai, H., Huang, S., Wu, G., Yang, D., Wang, S., Liao, J., Zhang, Y., Liu, M., & Liu, L. (2026) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2026.116132
Designing reproducible large-language-model-assisted scientific analyses
Dunn, C. W., Schultz, D. T., & Musser, J. M. (2026) · Patterns · doi:10.1016/j.patter.2026.101644 · Cited by 1
Risk-Adaptive Fine-Grained Authorization via Lightweight-LLM Collaboration
Li, M., & Peng, J. (2026) · International Conference on Cyber Security, Artificial Intelligence and Digital Economy (CSAIDE) · doi:10.1145/3819836.3819847
Cao, J., & Zhao, X. (2026) · International Conference on Cyber Security, Artificial Intelligence and Digital Economy (CSAIDE) · doi:10.1145/3819836.3819974
“[13] Small sample experimental protocol: set using the N-way K-shot standard.”
Evaluation of Vanilla and RAG-Supported LLM Comprehension of the EU AI Act
Tsalera, E., Pagiatakis, G. K., & Papadakis, A. E. (2026) · Future Internet · doi:10.3390/fi18090469 · Cited by 1
Predicting User Engagement in Niche Cultural Domains:A Hybrid LDA-BERT Approach for Short Text
Xu, C., He, T., Lin, S., & Kong, Y. (2026) · International Conference on Machine Intelligence and Digital Applications (MIDA) · doi:10.1145/3801438.3801475
Güneş, Z., & Mergen, E. K. (2026) · Orthopaedic Journal of Sports Medicine · doi:10.1177/23259671261463933
“17,33 A recent study found that LLMs could create a manuscript acceptable for submission to an orthopaedic journal writing with fabricated results and references.”
Transforming large language models into medical specialists via knowledge injection
Kim, K., Song, J., Kim, D. Y., Nam, Y., Park, S., Shin, K., Jo, W., & Kim, N. (2026) · Cell Reports Medicine · doi:10.1016/j.xcrm.2026.103020
Ieva, S., Loconte, D., Loseto, G., Ruta, M., Scioscia, F., Marche, D., & Notarnicola, M. (2026) · Elsevier eBooks · doi:10.1016/b978-0-443-36664-2.00024-1
Reducing hypertension disparities: The potential role of patient-facing large language models
Allouch, F., Hassan, S., Alvarado, F., Asad, S. M. N., Parekh, D., Johnson, R., Payne, M. J., Gardner, D., Mattei, N., & Ferdinand, K. C. (2026) · American Journal of Preventive Cardiology · doi:10.1016/j.ajpc.2026.101813
Gutiérrez, J. A. A., Pineda, H. R., & Tasamá, A. V. (2026) · South Florida Journal of Development · doi:10.46932/sfjdv7n8-033
Savant, S., Deshpande, U. U., & D’Silva, J. (2026) · Discover Artificial Intelligence · doi:10.1007/s44163-026-01750-x
“However, performance remains strongly influenced by the multilingual pre-training distribution and the representation of the corpus data across languages [23–25, 28–30, 65–70].”
La Apropiación de la Inteligencia Artificial en Docentes Universitarios
Rocha, I. R. R., Márquez, A. M. G., Jesus, A. F. K. D., & Acosta, G. C. (2026) · Estudios y Perspectivas Revista Científica y Académica · doi:10.61384/r.c.a..v6i3.2069
Uncertainty calibration for hallucination detection via token dependency and sequence divergence
Qin, S., Zhou, L., Sun, L., & Wang, N. (2026) · Applied Soft Computing · doi:10.1016/j.asoc.2026.116357
Kaşarcı, İ. (2026) · International Journal of Current Educational Studies · doi:10.46328/ijces.315 · Cited by 1
Shi, Y., Zheng, Y., Wei, Y., Ji, X., & Ke, Z. (2026) · Creativity Research Journal · doi:10.1080/10400419.2026.2726242
Bausch, D., Keller, S., & Leyh, C. (2026) · HMD Praxis der Wirtschaftsinformatik · doi:10.1365/s40702-026-01321-9
Stabilizing Graph Explanations Under Spurious Correlations via Self-reflection
Cai, K., Xu, C., Peng, C., Wang, Y., Huang, J., Liang, Q., & Zheng, P. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-37657-2_29
Sakurai, T., Aizawa, D., Okawa, K., Kofuchi, R., Hirota, M., Gotouda, H., Yamamoto, T., & Ohkubo, C. (2026) · Journal of Medical Education and Curricular Development · doi:10.1177/23821205261483568
From Brownfield CLI to Intent Models: A Guardrailed Agentic Translation Workflow
Suvarna, N., Pappu, K., Sharma, S., & Ageze, B. (2026) · Conference on Local Computer Networks · doi:10.1109/lcn67947.2026.11660787
Kumar, N. A., & Kataria, V. (2026) · Journal of Hospitality and Tourism Insights · doi:10.1108/jhti-12-2025-1458
Adversarial Fact Editing and Active RAG Defenses for Robust Large Language Model Bash Execution
Adler, E., Seitz, C., Holz, L., Loevenich, J. F., & Lopes, R. R. F. (2026) · Conference on Local Computer Networks · doi:10.1109/lcn67947.2026.11660783
Çörekci, G., & Çörekci, A. M. (2026) · Physics Education · doi:10.1088/1361-6552/ae9bf4
Agrawal, K., Akshay, D. A., Kaushik, P., Jain, S., Sharma, G., & Yogarayan, S. (2026) · Machine Learning and Knowledge Extraction · doi:10.3390/make8090274
“Hallucination—the generation of content that is factually incorrect or not supported by the input—is a well-documented failure mode of generative language models [33].”
Muñoz-Blat, M., Vicente-Martínez, P., Soria‐Olivas, E., Sánchez-Montañés, M. A., García-Escrivà, M. Á., & William, E. (2026) · International Journal of Gastronomy and Food Science · doi:10.1016/j.ijgfs.2026.101641
Jung, H. S., & Lee, H. (2026) · International Journal of Human-Computer Interaction · doi:10.1080/10447318.2026.2719118
Embodied intelligence machine tools: concepts, architecture and key technologies
Chen, Y., Yao, Z., & Peng, C. (2026) · The International Journal of Advanced Manufacturing Technology · doi:10.1007/s00170-026-18993-8
Borisov, I. (2026) · Human Resource Development Review · doi:10.1177/15344843261480993
Entropy-Guided Telemetry Compression for LLM-Based RCA: An Industrial Pilot on Log Data
García-Hernández, C. F., Abdali, A., Wu, G., Pang, Z., Shen, F., Zhang, Y., & Wang, M. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-37685-5_51
Song, H. (2026) · Computers & Chemical Engineering · doi:10.1016/j.compchemeng.2026.109889 · Cited by 1
A systematic research synthesis of generative AI literacy in English language education
Zhang, Y., Qiu, X., & Wang, J. (2026) · Discover Artificial Intelligence · doi:10.1007/s44163-026-02178-z · Cited by 7
“Hallucination and factual unreliability are well-documented concerns in the system itself [2–9], while delegation, privacy, bias, and academic-integrity risks require users’ pedagogical and ethical judgement [2, 10–11].”
Rajasekaran, S. K., Ali, A., & Devasia, C. (2026) · ITM Web of Conferences · doi:10.1051/itmconf/20269002010
“Natural language generation was radically altered by subsequent developments in transformer-based systems that employed self-attention to capture long-range linkages [9].”
Giraldo, M., & Sacchi, F. (2026) · Healthcare · doi:10.3390/healthcare14182900 · Cited by 3
Cheonghoon, L., & Jisoo, S. (2026) · Journal of the Korean Institute of Landscape Architecture · doi:10.9715/kila.2026.54.4.196
Trustworthy LLM–GNN systems: a systematic review and multi-axis taxonomy
Xue, R., He, F., Deng, H., Wang, M., & Zhang, Z. (2026) · Applied Intelligence · doi:10.1007/s10489-026-07458-w
“The defining LLM-specific reliability risk is hallucination: content that is inconsistent with, or unsupported by, the available evidence [33–35].”
Peng, Y., Liu, Y., Wang, Q., Zhang, L., Xie, C., Zhang, Y., & Mao, Z. (2026) · ACM Transactions on Knowledge Discovery from Data · doi:10.1145/3837071
Gualo, F., Ayuso, Y., Caballero, I., & Piattini, M. G. (2026) · SoftwareX · doi:10.1016/j.softx.2026.103012
Dong, K., Wang, R., Yuan, F., An, B., & Zhang, X. (2026) · IEEE Access · doi:10.1109/access.2026.3730896
“Directly sending raw logs to an LLM can therefore introduce noise and increase hallucination risk [18].”
ScholarGuard-AI An Intelligent Framework for Research Paper Generation and Authenticity Detection
Karthik, S., Kumar, T., Kumari, S., & Kumar, P. (2026) · ITM Web of Conferences · doi:10.1051/itmconf/20269002013
“Another way to check for research integrity is to look at the citations [7].”
EviGuard: Machine-Verifiable Evidence Grounding for LLM-Based Industrial Incident Reasoning
Zhou, H., Lei, H., & Yang, M. (2026) · Applied Sciences · doi:10.3390/app16188925
Why we believe chatbots: trust calibration as a design problem
Jaidka, K., & Cai, M. (2026) · Frontiers in Psychology · doi:10.3389/fpsyg.2026.1935527 · Cited by 4
“Beyond stylistic tropes, chatbots have been reported to produce fabricated but convincingly formatted evidence, including medical references (Alkaissi and McFarlane, 2023) and legal citations (Dahl et al., 2024), and hallucination is documented across model families (Ji et al., 2023).”
Linnes, C., Ronzoni, G., Agrusa, J. F., GEORGE, B. P., & Lema, J. D. (2026) · Tourism and Hospitality · doi:10.3390/tourhosp7090293
Metadata
- Authors
- Ji, Z., Lee, N., Frieske, R., Yu, T., Su, D., Xu, Y., Ishii, E., Bang, Y. J., Madotto, A., & Fung, P.
- Affiliations
- Hong Kong University of Science and Technology (HK)
- ORCID
- Venue
- ACM Computing Surveys
- Published
- 2023-03-03
- Volume
- 55
- Issue
- 12
- Pages
- 1-38
- ISSN
- 0360-0300 (ISSN-L)
- 1557-7341
- OpenAlex
- W4309674289
- Semantic Scholar
- 3def68bd0f856886d34272840a7f81588f2bc082
- Cited by
- 5,025
- References
- 111
- Citations by year
- 2026: 2296 · 2025: 1441 · 2024: 817 · 2023: 453
- FWCI
- 224.36
- Type
- journal article
- Language
- en
- Open access
- green
- arxiv.org
- Locations
- ACM Computing Surveys (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- Rare & Special e-Zone (The Hong Kong University of Science and Technology) (submittedVersion)
- arXiv (Cornell University) (cc-by)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Topic Modeling
- Physical Sciences › Computer Science › Artificial Intelligence › Natural Language Processing Techniques
- Physical Sciences › Computer Science › Artificial Intelligence › Speech and dialogue systems
- Keywords
- hallucinations, natural language generation, hallucination mitigation, abstractive summarization, dialogue generation, generative question answering, data-to-text generation, machine translation, large language models, transformer models
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.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-09-30, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-09-30 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol).
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
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.




