---
title: "ARAFA: An LLM-Generated Arabic Fact-Checking Dataset — Khalil 2026"
description: "Earlier Arabic sets such as AraFacts and AraStance are small and mostly limited to news or posts on X. The authors build ARAFA, a Modern… 45 references."
authors: ["Christophe Khalil", "Shady Elbassuoni", "Rida Assaf"]
year: 2026
venue: "arXiv:2609.25833"
source: "https://arxiv.org/abs/2609.25833"
doi: "10.48550/arxiv.2609.25833"
arxiv: "2609.25833"
licence: "CC BY-NC-SA 4.0 (arXiv)"
licence_url: "http://creativecommons.org/licenses/by-nc-sa/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/khalil-2026-arafa-llm-generated-arabic-fact/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/khalil-2026-arafa-llm-generated-arabic-fact/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2026](https://md.research.elamur.ai/free/en/papers/2026/page/4/) / ARAFA: An LLM-Generated Arabic Fact-Checking Dataset

# ARAFA: An LLM-Generated Arabic Fact-Checking Dataset

**Authors:** Christophe Khalil, Shady Elbassuoni and Rida Assaf · **Year:** 2026 · **Venue:** arXiv:2609.25833 · **Licence:** [CC BY-NC-SA 4.0 (arXiv)](http://creativecommons.org/licenses/by-nc-sa/4.0/)

**Links:** [Publisher's page](https://arxiv.org/abs/2609.25833) · [PDF](https://arxiv.org/pdf/2609.25833) · [DOI 10.48550/arxiv.2609.25833](https://doi.org/10.48550/arxiv.2609.25833) · [arXiv 2609.25833](https://arxiv.org/abs/2609.25833) · [OpenAlex](https://openalex.org/W7214086152)

**Citations:** 0 citations (OpenAlex, 2026-10-14)

## Summary
Arabic automatic fact-checking has lacked a large, general-domain training resource. Earlier Arabic sets such as AraFacts and AraStance are small and mostly limited to news or posts on X. The authors build ARAFA, a Modern Standard Arabic dataset produced almost entirely by LLMs, using a three-stage pipeline:
- GPT-4o reads Arabic Wikipedia pages and writes true claims together with the passages that back them.
- GPT-4o then alters some of these claims into counterfactual versions that their evidence contradicts.
- Claude Sonnet 3.5 independently gives each claim-evidence pair one of three labels: refuted, supported, or not enough information (NEI).

The result is 181,976 labeled pairs covering history, science, society, biographies and more. Checked against expert annotations, the automatic validator was right on 86% of supported and 88% of refuted pairs, which is close to how well human annotators agreed with each other. Four fine-tuned open models reach up to 77% macro F1. All of them struggle with the NEI class. The work matters as the largest open Arabic fact-checking benchmark and as a reusable recipe for other low-resource languages.

## Key points
- Label distribution: 111,303 supported, 42,109 refuted and 28,564 NEI pairs. Data is on Zenodo and code is at github.com/chriskhalil/ARAFA, including all prompts.
- Human evaluation design:
  - Two experts labeled a sample of 200 supported and 200 refuted pairs.
  - Four further annotators labeled the same sample. Their mutual agreement was κ = 0.89 (supported) and 0.94 (refuted), and their agreement with the experts was κ = 0.86 and 0.84.
- Four fine-tuned models were compared. The generative ones were trained with LoRA.
  - AraModernBert-Base-V1.0 had the best macro F1 (77%).
  - AraBERTv2 and Qwen2.5-7B tied for the best accuracy (82%).
  - Llama-3.1-8B was completed with the others in the comparison.
- Per-class F1:
  - Supported: 85–89% across models (Qwen2.5-7B best at 89%).
  - Refuted: 86–93% (Llama-3.1-8B best at 93%).
  - NEI: only 47–52%.
- Stated limitations:
  - The data comes from Wikipedia only and covers MSA only, with no dialects.
  - LLM artifacts or hallucinations may carry over into the data.
  - Claim complexity is capped by what the generating model can handle in morphologically rich Arabic.

## Interesting ideas
- Generate positives with one LLM and have a different LLM family relabel them, so a single model's blind spots do not silently become gold labels.
- Make hard negatives by deliberately altering true statements (for example numbers, dates or roles) and pairing them with evidence that contradicts them.
- Let the validator assign a third outcome when evidence is merely related but not decisive, rather than forcing a yes/no verdict.
- Judge an automatic labeler against a small expert-labeled sample, and use agreement among non-expert humans as the realistic ceiling.
- Break results down by class to expose a shared weakness (here NEI) that a single headline score would hide.

## For Mova

How these ideas could serve Mova, our small open language model — its files, stages and tests:

- **I1** (learn): our LLM teacher tiers could pair a generator teacher with an independent checker teacher before any silver example enters training.
- **I2** (learn): such perturbed sentences could serve as negative controls, testing whether the world model actually catches a contradicted state instead of matching surface overlap.
- **I3** (answer): our readers could explicitly output "the text does not say", grounded in a missing entity, state or event in the world of the text, rather than guessing.
- **I4** (learn): when validating silver parses or frames, we could report teacher-vs-gold accuracy next to human-vs-human agreement, so we know when the automatic labeler is good enough.
- **I5** (answer): our school-test and fable QA evaluations should split scores by answer type, especially unanswerable questions, to locate where the reasoning chain breaks.

*Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.*

Fields: [Linguistics & languages](https://md.research.elamur.ai/free/en/shelves/linguistics/)

Topics: [🐪 Semitic languages](https://md.research.elamur.ai/free/en/topics/semitic-languages/)

## Institutions

- [American University of Beirut](https://md.research.elamur.ai/free/en/universities/american-university-of-beirut/) — University, Beirut, Lebanon · 11 works · 1,412 citations

## Cite

Cite as: Christophe Khalil, Shady Elbassuoni and Rida Assaf. 2026. ARAFA: An LLM-Generated Arabic Fact-Checking Dataset. arXiv:2609.25833.

The paper itself, not this page. BibTeX from the library's record of the work.

```bibtex
@misc{khalil2026arafa,
  title = {{ARAFA: An LLM-Generated Arabic Fact-Checking Dataset}},
  author = {Christophe Khalil, Shady Elbassuoni and Rida Assaf},
  year = {2026},
  doi = {10.48550/arxiv.2609.25833},
  eprint = {2609.25833},
  archivePrefix = {arXiv},
  url = {https://doi.org/10.48550/arxiv.2609.25833}
}
```

- [← previous AraDynFact: Dynamic Evaluation of Factual Knowledge in Arabic](https://md.research.elamur.ai/free/en/papers/iacobacci-2026-aradynfact-dynamic-evaluation-factual-knowledge/)
- [next → AraMIP: Extending MIPVU Towards Metaphor Identification in Arabic](https://md.research.elamur.ai/free/en/papers/marathe-2026-aramip-extending-mipvu-towards-metaphor/)
- [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/4/)

## References

45 references, 15 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Tariq Alhindi, Amal Alabdulkarim, Ali Alshehri, Muhammad Abdul-Mageed, and Preslav Nakov. 2021. Arastance: A multi-country and multi-domain dataset of arabic stance detection for fact checking. arXiv preprint arXiv:2104.13559. [↗](https://arxiv.org/abs/2104.13559) [our card](https://md.research.elamur.ai/free/en/papers/alhindi-2021-arastance-multi-country-multi-domain/)
2. Zien Sheikh Ali, Watheq Mansour, Tamer Elsayed, and Abdulaziz Al-Ali. 2021. Arafacts: The first large arabic dataset of naturally-occurring professionally-verified claims. Association for Computational Linguistics (ACL). [find](https://scholar.google.com/scholar?q=Zien+Sheikh+Ali%2C+Watheq+Mansour%2C+Tamer+Elsayed%2C+and+Abdulaziz+Al%2DAli%2E+2021%2E+Arafacts%3A+The+first+large+arabic+dataset+of+naturally%2Doccurring+professionally%2Dverified+claims%2E+Association+for+Computationa)
3. Saud Althabiti, Mohammad Ammar Alsalka, and Eric Atwell. 2024. Ta’keed: The first generative fact-checking system for arabic claims. arXiv preprint arXiv:2401.14067. [↗](https://arxiv.org/abs/2401.14067) [our card](https://md.research.elamur.ai/free/en/papers/althabiti-2024-ta-keed-first-generative-fact/)
4. Rami Aly, Zhijiang Guo, Michael Schlichtkrull, James Thorne, Andreas Vlachos, Christos Christodoulopoulos, Oana Cocarascu, and Arpit Mittal. 2021. Feverous: Fact extraction and verification over unstructured and structured information. arXiv preprint arXiv:2106.05707. [↗](https://arxiv.org/abs/2106.05707)
5. Jacopo Amidei, Paul Piwek, and Alistair Willis. 2019. The use of rating and likert scales in natural language generation human evaluation tasks: A review and some recommendations. [find](https://scholar.google.com/scholar?q=Jacopo+Amidei%2C+Paul+Piwek%2C+and+Alistair+Willis%2E+2019%2E+The+use+of+rating+and+likert+scales+in+natural+language+generation+human+evaluation+tasks%3A+A+review+and+some+recommendations%2E)
6. Wissam Antoun, Fady Baly, and Hazem Hajj. 2020. Arabert: Transformer-based model for arabic language understanding. arXiv preprint arXiv:2003.00104. [↗](https://arxiv.org/abs/2003.00104) [our card](https://md.research.elamur.ai/free/en/papers/antoun-2020-arabert-transformer-based-model-arabic/)
7. Catherine Arnett and Benjamin K Bergen. 2024. Why do language models perform worse for morphologically complex languages? arXiv preprint arXiv:2411.14198. [↗](https://arxiv.org/abs/2411.14198)
8. Emily M Bender, Timnit Gebru, Angelina McMillan-Major, and Shmargaret Shmitchell. 2021. On the dangers of stochastic parrots: Can language models be too big? In Proceedings of the 2021 ACM conference on fairness, accountability, and transparency, pages 610–623. [↗](https://doi.org/10.1145/3442188.3445922)
9. Tom Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared D Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, et al. 2020. Language models are few-shot learners. Advances in neural information processing systems, 33:1877–1901. [our card](https://md.research.elamur.ai/free/en/papers/brown-2020-language-models-few-shot-learners/)
10. Tsun-Hin Cheung and Kin-Man Lam. 2023. Factllama: Optimizing instruction-following language models with external knowledge for automated fact-checking. In 2023 Asia Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC), pages 846–853. IEEE. [find](https://scholar.google.com/scholar?q=Tsun%2DHin+Cheung+and+Kin%2DMan+Lam%2E+2023%2E+Factllama%3A+Optimizing+instruction%2Dfollowing+language+models+with+external+knowledge+for+automated+fact%2Dchecking%2E+In+2023+Asia+Pacific+Signal+and+Information+Proc)
11. Yi-Ling Chung, Aurora Cobo, and Pablo Serna. 2025. Beyond translation: Llm-based data generation for multilingual fact-checking. arXiv preprint arXiv:2502.15419. [↗](https://arxiv.org/abs/2502.15419)
12. Yufeng Du, Minyang Tian, Srikanth Ronanki, Subendhu Rongali, Sravan Bodapati, Aram Galstyan, Azton Wells, Roy Schwartz, Eliu A Huerta, and Hao Peng. 2025. Context length alone hurts llm performance despite perfect retrieval. arXiv preprint arXiv:2510.05381. [↗](https://arxiv.org/abs/2510.05381)
13. Omar Elshehy, Omer Nacar, Abdelbasset Djamai, Muhammed Ragab, Khloud AL Jallad, and Mona Abdelazim. 2026. Aramodernbert: Transtokenized initialization and long-context encoder modeling for arabic. In Proceedings of the 2nd Workshop on NLP for Languages Using Arabic Script, pages 313–321. [find](https://scholar.google.com/scholar?q=Omar+Elshehy%2C+Omer+Nacar%2C+Abdelbasset+Djamai%2C+Muhammed+Ragab%2C+Khloud+AL+Jallad%2C+and+Mona+Abdelazim%2E+2026%2E+Aramodernbert%3A+Transtokenized+initialization+and+long%2Dcontext+encoder+modeling+for+arabic%2E+In+)
14. Yoav Goldberg. 2019. Assessing bert’s syntactic abilities. arXiv preprint arXiv:1901.05287. [↗](https://arxiv.org/abs/1901.05287) [our card](https://md.research.elamur.ai/free/en/papers/goldberg-2019-assessing-bert-s-syntactic-abilities/)
15. Aaron Grattafiori, Abhimanyu Dubey, Abhinav Jauhri, Abhinav Pandey, Abhishek Kadian, Ahmad Al-Dahle, Aiesha Letman, Akhil Mathur, Alan Schelten, Alex Vaughan, et al. 2024. The llama 3 herd of models. arXiv preprint arXiv:2407.21783. [↗](https://arxiv.org/abs/2407.21783)
16. Andreas Hanselowski, Christian Stab, Claudia Schulz, Zile Li, and Iryna Gurevych. 2019. A richly annotated corpus for different tasks in automated fact-checking. arXiv preprint arXiv:1911.01214. [↗](https://arxiv.org/abs/1911.01214)
17. Fatima Haouari, Tamer Elsayed, and Reem Suwaileh. 2024. Aured: enabling arabic rumor verification using evidence from authorities over twitter. In Proceedings of The Second Arabic Natural Language Processing Conference, pages 27–41. [find](https://scholar.google.com/scholar?q=Fatima+Haouari%2C+Tamer+Elsayed%2C+and+Reem+Suwaileh%2E+2024%2E+Aured%3A+enabling+arabic+rumor+verification+using+evidence+from+authorities+over+twitter%2E+In+Proceedings+of+The+Second+Arabic+Natural+Language+Pro)
18. David M Howcroft, Anja Belz, Miruna Clinciu, Dimitra Gkatzia, Sadid A Hasan, Saad Mahamood, Simon Mille, Emiel Van Miltenburg, Sashank Santhanam, and Verena Rieser. 2020. Twenty years of confusion in human evaluation: Nlg needs evaluation sheets and standardised definitions. In Proceedings of the 13th international conference on natural language generation, pages 169–182. [find](https://scholar.google.com/scholar?q=David+M+Howcroft%2C+Anja+Belz%2C+Miruna+Clinciu%2C+Dimitra+Gkatzia%2C+Sadid+A+Hasan%2C+Saad+Mahamood%2C+Simon+Mille%2C+Emiel+Van+Miltenburg%2C+Sashank+Santhanam%2C+and+Verena+Rieser%2E+2020%2E+Twenty+years+of+confusion+in+)
19. Cheng-Ping Hsieh, Simeng Sun, Samuel Kriman, Shantanu Acharya, Dima Rekesh, Fei Jia, Yang Zhang, and Boris Ginsburg. 2024. Ruler: What’s the real context size of your long-context language models? arXiv preprint arXiv:2404.06654. [↗](https://arxiv.org/abs/2404.06654) [our card](https://md.research.elamur.ai/free/en/papers/hsieh-2024-ruler-s-real-context-size/)
20. Edward J Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, Weizhu Chen, et al. 2022a. Lora: Low-rank adaptation of large language models. ICLR, 1(2):3. [our card](https://md.research.elamur.ai/free/en/papers/hu-2021-lora-low-rank-adaptation-large/)
21. Xuming Hu, Zhijiang Guo, GuanYu Wu, Aiwei Liu, Lijie Wen, and Philip S Yu. 2022b. Chef: A pilot chinese dataset for evidence-based fact-checking. arXiv preprint arXiv:2206.11863. [↗](https://arxiv.org/abs/2206.11863)
22. Aminul Huq, Mst Pervin, et al. 2020. Adversarial attacks and defense on texts: A survey. arXiv preprint arXiv:2005.14108. [↗](https://arxiv.org/abs/2005.14108)
23. Vincent Jung and Lonneke van der Plas. 2024. Understanding the effects of language-specific class imbalance in multilingual fine-tuning. arXiv preprint arXiv:2402.13016. [↗](https://arxiv.org/abs/2402.13016)
24. Daniel Kahneman. 2011. Thinking, fast and slow. Farrar, Straus and Giroux. [find](https://scholar.google.com/scholar?q=Daniel+Kahneman%2E+2011%2E+Thinking%2C+fast+and+slow%2E+Farrar%2C+Straus+and+Giroux%2E)
25. Neema Kotonya and Francesca Toni. 2020. Explainable automated fact-checking for public health claims. arXiv preprint arXiv:2010.09926. [↗](https://arxiv.org/abs/2010.09926) [our card](https://md.research.elamur.ai/free/en/papers/kotonya-2020-explainable-automated-fact-checking-public/)
26. Nayeon Lee, Belinda Z Li, Sinong Wang, Wen-tau Yih, Hao Ma, and Madian Khabsa. 2020. Language models as fact checkers? arXiv preprint arXiv:2006.04102. [↗](https://arxiv.org/abs/2006.04102)
27. Ying-Jia Lin, Chun-Yi Lin, Chia-Jen Yeh, Yi-Ting Li, Yun-Yu Hu, Chih-Hao Hsu, Mei-Feng Lee, and Hung-Yu Kao. 2024. Cfever: A chinese fact extraction and verification dataset. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, pages 18626–18634. [find](https://scholar.google.com/scholar?q=Ying%2DJia+Lin%2C+Chun%2DYi+Lin%2C+Chia%2DJen+Yeh%2C+Yi%2DTing+Li%2C+Yun%2DYu+Hu%2C+Chih%2DHao+Hsu%2C+Mei%2DFeng+Lee%2C+and+Hung%2DYu+Kao%2E+2024%2E+Cfever%3A+A+chinese+fact+extraction+and+verification+dataset%2E+In+Proceedings+of+the+AAA)
28. Zhenghao Liu, Chenyan Xiong, Maosong Sun, and Zhiyuan Liu. 2019. Fine-grained fact verification with kernel graph attention network. arXiv preprint arXiv:1910.09796. [↗](https://arxiv.org/abs/1910.09796) [our card](https://md.research.elamur.ai/free/en/papers/liu-2020-fine-grained-fact-verification-kernel/)
29. Ilya Loshchilov and Frank Hutter. 2017. Decoupled weight decay regularization. arXiv preprint arXiv:1711.05101. [↗](https://arxiv.org/abs/1711.05101)
30. Huanhuan Ma, Weizhi Xu, Yifan Wei, Liuji Chen, Liang Wang, Qiang Liu, and Shu Wu. 2024. Ex-fever: A dataset for multi-hop explainable fact verification. In Findings of the Association for Computational Linguistics: ACL 2024, pages 9340–9353. [find](https://scholar.google.com/scholar?q=Huanhuan+Ma%2C+Weizhi+Xu%2C+Yifan+Wei%2C+Liuji+Chen%2C+Liang+Wang%2C+Qiang+Liu%2C+and+Shu+Wu%2E+2024%2E+Ex%2Dfever%3A+A+dataset+for+multi%2Dhop+explainable+fact+verification%2E+In+Findings+of+the+Association+for+Computationa)
31. Leland McInnes, John Healy, and James Melville. 2018. Umap: Uniform manifold approximation and projection for dimension reduction. arXiv preprint arXiv:1802.03426. [↗](https://arxiv.org/abs/1802.03426) [our card](https://md.research.elamur.ai/free/en/papers/mcinnes-2018-umap-uniform-manifold-approximation-projection/)
32. Yixin Nie, Haonan Chen, and Mohit Bansal. 2019. Combining fact extraction and verification with neural semantic matching networks. In Proceedings of the AAAI conference on artificial intelligence, volume 33, pages 6859–6866. [find](https://scholar.google.com/scholar?q=Yixin+Nie%2C+Haonan+Chen%2C+and+Mohit+Bansal%2E+2019%2E+Combining+fact+extraction+and+verification+with+neural+semantic+matching+networks%2E+In+Proceedings+of+the+AAAI+conference+on+artificial+intelligence%2C+vol)
33. Liangming Pan, Wenhu Chen, Wenhan Xiong, Min-Yen Kan, and William Yang Wang. 2021. Zero-shot fact verification by claim generation. arXiv preprint arXiv:2105.14682. [↗](https://arxiv.org/abs/2105.14682)
34. Marco Tulio Ribeiro, Sameer Singh, and Carlos Guestrin. 2018. Semantically equivalent adversarial rules for debugging nlp models. In Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (volume 1: long papers), pages 856–865. [find](https://scholar.google.com/scholar?q=Marco+Tulio+Ribeiro%2C+Sameer+Singh%2C+and+Carlos+Guestrin%2E+2018%2E+Semantically+equivalent+adversarial+rules+for+debugging+nlp+models%2E+In+Proceedings+of+the+56th+Annual+Meeting+of+the+Association+for+Compu)
35. Dominik Stammbach and Elliott Ash. 2020. e-fever: Explanations and summaries for automated fact checking. Proceedings of the 2020 Truth and Trust Online (TTO 2020), pages 32–43. [find](https://scholar.google.com/scholar?q=Dominik+Stammbach+and+Elliott+Ash%2E+2020%2E+e%2Dfever%3A+Explanations+and+summaries+for+automated+fact+checking%2E+Proceedings+of+the+2020+Truth+and+Trust+Online+%28TTO+2020%29%2C+pages+32%E2%80%9343%2E)
36. Qwen Team. 2024. Qwen2 technical report. arXiv preprint arXiv:2412.15115. [↗](https://arxiv.org/abs/2412.15115)
37. James Thorne and Andreas Vlachos. 2018. Automated fact checking: Task formulations, methods and future directions. arXiv preprint arXiv:1806.07687. [↗](https://arxiv.org/abs/1806.07687) [our card](https://md.research.elamur.ai/free/en/papers/thorne-2018-automated-fact-checking-task-formulations/)
38. James Thorne, Andreas Vlachos, Christos Christodoulopoulos, and Arpit Mittal. 2018. Fever: a large-scale dataset for fact extraction and verification. arXiv preprint arXiv:1803.05355. [↗](https://arxiv.org/abs/1803.05355) [our card](https://md.research.elamur.ai/free/en/papers/thorne-2018-fever-large-scale-dataset-fact/)
39. Soroush Vosoughi, Deb Roy, and Sinan Aral. 2018. The spread of true and false news online. science, 359(6380):1146–1151. [find](https://scholar.google.com/scholar?q=Soroush+Vosoughi%2C+Deb+Roy%2C+and+Sinan+Aral%2E+2018%2E+The+spread+of+true+and+false+news+online%2E+science%2C+359%286380%29%3A1146%E2%80%931151%2E)
40. David Wadden, Kyle Lo, Bailey Kuehl, Arman Cohan, Iz Beltagy, Lucy Lu Wang, and Hannaneh Hajishirzi. 2022. Scifact-open: Towards open-domain scientific claim verification. arXiv preprint arXiv:2210.13777. [↗](https://arxiv.org/abs/2210.13777)
41. Haoran Wang and Kai Shu. 2023. Explainable claim verification via knowledge-grounded reasoning with large language models. arXiv preprint arXiv:2310.05253. [↗](https://arxiv.org/abs/2310.05253)
42. Liang Wang, Nan Yang, Xiaolong Huang, Binxing Jiao, Linjun Yang, Daxin Jiang, Rangan Majumder, and Furu Wei. 2022. Text embeddings by weakly-supervised contrastive pre-training. arXiv preprint arXiv:2212.03533. [↗](https://arxiv.org/abs/2212.03533) [our card](https://md.research.elamur.ai/free/en/papers/wang-2022-text-embeddings-weakly-supervised-contrastive/)
43. Xuan Zhang and Wei Gao. 2023. Towards llm-based fact verification on news claims with a hierarchical step-by-step prompting method. arXiv preprint arXiv:2310.00305. [↗](https://arxiv.org/abs/2310.00305)
44. Wayne Xin Zhao, Kun Zhou, Junyi Li, Tianyi Tang, Xiaolei Wang, Yupeng Hou, Yingqian Min, Beichen Zhang, Junjie Zhang, Zican Dong, et al. 2023. A survey of large language models. arXiv preprint arXiv:2303.18223. [↗](https://arxiv.org/abs/2303.18223) [our card](https://md.research.elamur.ai/free/en/papers/zhao-2023-survey-large-language-models/)
45. Jie Zhou, Xu Han, Cheng Yang, Zhiyuan Liu, Lifeng Wang, Changcheng Li, and Maosong Sun. 2019. Gear: Graph-based evidence aggregating and reasoning for fact verification. arXiv preprint arXiv:1908.01843. [↗](https://arxiv.org/abs/1908.01843) [our card](https://md.research.elamur.ai/free/en/papers/zhou-2019-gear-graph-based-evidence-aggregating/)