---
title: "GAVEL — paper by Xu et al. (2026) — summary and ideas"
description: "Fact-checking systems built on LLMs can give fluent justifications that do not trace back to the exact evidence. 27 references."
authors: ["Ruoyu Xu", "Gaoxiang Li", "Victor S. Sheng"]
year: 2026
venue: "Findings of the Association for Computational Linguistics: ACL 2026"
source: "https://aclanthology.org/2026.findings-acl.1789/"
doi: "10.18653/v1/2026.findings-acl.1789"
licence: "CC BY 4.0 (ACL Anthology)"
licence_url: "https://creativecommons.org/licenses/by/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/xu-2026-gavel-evidence-contract-debate-mechanized/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/xu-2026-gavel-evidence-contract-debate-mechanized/)

[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/13/) / GAVEL: Evidence-Contract Debate with Mechanized Scrutiny for Provenance-Grounded Fact-Checking

# GAVEL: Evidence-Contract Debate with Mechanized Scrutiny for Provenance-Grounded Fact-Checking

**Authors:** Ruoyu Xu, Gaoxiang Li and Victor S. Sheng · **Year:** 2026 · **Venue:** [Findings of the Association for Computational Linguistics: ACL 2026](https://md.research.elamur.ai/free/en/venues/findings/) ([Findings of the Association for Computational Linguistics 2026](https://md.research.elamur.ai/free/en/venues/findings/2026/)) · **Publisher:** [Association for Computational Linguistics](https://md.research.elamur.ai/free/en/publishers/association-for-computational-linguistics/) · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/2026.findings-acl.1789/) · [PDF](https://aclanthology.org/2026.findings-acl.1789.pdf) · [DOI 10.18653/v1/2026.findings-acl.1789](https://doi.org/10.18653/v1/2026.findings-acl.1789) · [OpenAlex](https://openalex.org/W7166808616)

**Citations:** 1 citations (OpenAlex, 2026-10-14)

## Summary
Fact-checking systems built on LLMs can give fluent justifications that do not trace back to the exact evidence. This is a serious weakness on benchmarks such as FEVEROUS and HOVER. Both score a prediction as correct only when the label is right and the returned evidence is correct and complete. That evidence can be individual sentences, specific table cells, or chains spanning several documents.

GAVEL is a multi-agent debate framework that applies provenance constraints at every step of inference. Two opposing debaters must split their arguments into atomic subclaims, and each subclaim must be tied to explicit evidence identifiers. This is the paper's "Evidence Contract." A neutral Scrutinizer then audits the arguments for logical gaps. It also runs a deterministic check that every cited ID and quoted span actually appears in the retrieved context. A Judge chooses a sufficient evidence set and issues the verdict.

In open-book experiments on Wikipedia, GAVEL beats ProgramFC, PACAR and LoCal on metrics that require both the right label and the right evidence. Ablations show that evidence binding and mechanized citation checks drive most of the gains. The work points toward fact-checkers whose outputs can be audited.

## Key points
- The framework uses four role-fixed agents (Affirmative, Negative, Scrutinizer, Judge), all running gpt-3.5-turbo at temperature 0. Debate is capped at three rounds, with early stopping once no new violations are found.
- Evidence comes in two forms:
  - sentence units ⟨page, sentence id⟩, used on both datasets;
  - table-cell units ⟨page, table, row, column⟩, used on FEVEROUS only.
- Label sets differ by benchmark: FEVEROUS has three labels (SUPPORTS / REFUTES / NOT ENOUGH INFO), and HOVER uses the binary SUPPORTED / NOT SUPPORTED setting.
- Average calls per claim:

  | System | LLM calls | Retrieval calls |
  |---|---|---|
  | GAVEL | 18 | 6 |
  | LoCal | 14 | 6 |
  | PACAR | 10 | 5 |
  | ProgramFC (N=5) | 30 | 15 |

  The deterministic checker adds almost no overhead.
- Remaining errors include ambiguous table headers (right row, wrong column) and missing intermediate pages in HOVER chains. The system also stays dependent on retrieval quality, and its checks do not verify numerical reasoning semantically.

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

## Institutions

Also on the byline: [Texas Tech University](https://ror.org/0405mnx93).

📚 Shelves (7): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 18,614 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 17,965 works · [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 17,450 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 17,296 works · [👁️ Computer Vision and Pattern Recognition](https://md.research.elamur.ai/free/en/shelves/computer-science/computer-vision-and-pattern-recognition/) · 3,486 works · [🤝 Social Sciences](https://md.research.elamur.ai/free/en/shelves/social-sciences/) · 1,954 works · [🤝 Sociology and Political Science](https://md.research.elamur.ai/free/en/shelves/social-sciences/sociology-and-political-science/) · 626 works

## Cite

Cite as: Ruoyu Xu, Gaoxiang Li and Victor S. Sheng. 2026. GAVEL: Evidence-Contract Debate with Mechanized Scrutiny for Provenance-Grounded Fact-Checking. Findings of the Association for Computational Linguistics: ACL 2026.

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

```bibtex
@inproceedings{xu2026gavel,
  title = {{GAVEL: Evidence-Contract Debate with Mechanized Scrutiny for Provenance-Grounded Fact-Checking}},
  author = {Ruoyu Xu, Gaoxiang Li and Victor S. Sheng},
  booktitle = {Findings of the Association for Computational Linguistics: ACL 2026},
  year = {2026},
  doi = {10.18653/v1/2026.findings-acl.1789},
  url = {https://doi.org/10.18653/v1/2026.findings-acl.1789}
}
```

- [← previous GAPO: Robust Advantage Estimation for Real-World Code LLMs](https://md.research.elamur.ai/free/en/papers/zhang-2026-gapo-robust-advantage-estimation-real/)
- [next → GenAIus at SemEval-2026 Task 8: Beyond Retrieval with Relevance-Aware RAG for Faithful Multi-Turn Generation](https://md.research.elamur.ai/free/en/papers/yeniterzi-2026-genaius-semeval-2026-task-8/)
- [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/13/)

## References

27 references, 10 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Esma Aïmeur, Sabrine Amri, and Gilles Brassard. 2023. Fake news, disinformation and misinformation in social media: a review. Social Network Analysis and Mining, 13(1):30. [find](https://scholar.google.com/scholar?q=Esma+A%C3%AFmeur%2C+Sabrine+Amri%2C+and+Gilles+Brassard%2E+2023%2E+Fake+news%2C+disinformation+and+misinformation+in+social+media%3A+a+review%2E+Social+Network+Analysis+and+Mining%2C+13%281%29%3A30%2E)
2. Rami Aly, Zhijiang Guo, Michael Sejr Schlichtkrull, James Thorne, Andreas Vlachos, Christos Christodoulopoulos, Oana Cocarascu, and Arpit Mittal. 2021. FEVEROUS: Fact extraction and VERification over unstructured and structured information. In Thirty-fifth Conference on Neural Information Processing… [find](https://scholar.google.com/scholar?q=Rami+Aly%2C+Zhijiang+Guo%2C+Michael+Sejr+Schlichtkrull%2C+James+Thorne%2C+Andreas+Vlachos%2C+Christos+Christodoulopoulos%2C+Oana+Cocarascu%2C+and+Arpit+Mittal%2E+2021%2E+FEVEROUS%3A+Fact+extraction+and+VERification+over+)
3. Giorgio Barnabò, Federico Siciliano, Carlos Castillo, Stefano Leonardi, Preslav Nakov, Giovanni Da San Martino, and Fabrizio Silvestri. 2023. Deep active learning for misinformation detection using geometric deep learning. Online Social Networks and Media, 33:100244. [find](https://scholar.google.com/scholar?q=Giorgio+Barnab%C3%B2%2C+Federico+Siciliano%2C+Carlos+Castillo%2C+Stefano+Leonardi%2C+Preslav+Nakov%2C+Giovanni+Da+San+Martino%2C+and+Fabrizio+Silvestri%2E+2023%2E+Deep+active+learning+for+misinformation+detection+using+ge)
4. Shehzaad Dhuliawala, Mojtaba Komeili, Jing Xu, Roberta Raileanu, Xian Li, Asli Celikyilmaz, and Jason Weston. 2024. [Chain-of-verification reduces hallucination in large language models](https://md.research.elamur.ai/free/en/papers/dhuliawala-2023-chain-verification-reduces-hallucination-large/). In Findings of the Association for Computational Linguistics: ACL 2024, pages 3563–3578, Bangkok, Thailand. Associ…
5. Miriam Fernandez and Harith Alani. 2018. Online misinformation: Challenges and future directions. In Companion proceedings of the the web conference 2018, pages 595–602. [find](https://scholar.google.com/scholar?q=Miriam+Fernandez+and+Harith+Alani%2E+2018%2E+Online+misinformation%3A+Challenges+and+future+directions%2E+In+Companion+proceedings+of+the+the+web+conference+2018%2C+pages+595%E2%80%93602%2E)
6. Yi Fung, Kung-Hsiang Huang, Preslav Nakov, and Heng Ji. 2022. The battlefront of combating misinformation and coping with media bias. In Proceedings of the 2nd Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics and the 12th International Joint Conference on Natur… [find](https://scholar.google.com/scholar?q=Yi+Fung%2C+Kung%2DHsiang+Huang%2C+Preslav+Nakov%2C+and+Heng+Ji%2E+2022%2E+The+battlefront+of+combating+misinformation+and+coping+with+media+bias%2E+In+Proceedings+of+the+2nd+Conference+of+the+Asia%2DPacific+Chapter+o)
7. Max Glockner, Yufang Hou, and Iryna Gurevych. 2022. Missing counter-evidence renders NLP fact-checking unrealistic for misinformation. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pages 5916–5936, Abu Dhabi, United Arab Emirates. Association for Computat… [find](https://scholar.google.com/scholar?q=Max+Glockner%2C+Yufang+Hou%2C+and+Iryna+Gurevych%2E+2022%2E+Missing+counter%2Devidence+renders+NLP+fact%2Dchecking+unrealistic+for+misinformation%2E+In+Proceedings+of+the+2022+Conference+on+Empirical+Methods+in+Nat)
8. Zhijiang Guo, Michael Schlichtkrull, and Andreas Vlachos. 2022. [A survey on automated fact-checking](https://md.research.elamur.ai/free/en/papers/guo-2022-survey-automated-fact-checking/). Transactions of the Association for Computational Linguistics, 10:178–206.
9. Linmei Hu, Siqi Wei, Ziwang Zhao, and Bin Wu. 2022. Deep learning for fake news detection: A comprehensive survey. AI Open, 3:133–155. [find](https://scholar.google.com/scholar?q=Linmei+Hu%2C+Siqi+Wei%2C+Ziwang+Zhao%2C+and+Bin+Wu%2E+2022%2E+Deep+learning+for+fake+news+detection%3A+A+comprehensive+survey%2E+AI+Open%2C+3%3A133%E2%80%93155%2E)
10. Kelvin Jiang, Ronak Pradeep, and Jimmy Lin. 2021. Exploring listwise evidence reasoning with t5 for fact verification. In Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 2: Shor… [find](https://scholar.google.com/scholar?q=Kelvin+Jiang%2C+Ronak+Pradeep%2C+and+Jimmy+Lin%2E+2021%2E+Exploring+listwise+evidence+reasoning+with+t5+for+fact+verification%2E+In+Proceedings+of+the+59th+Annual+Meeting+of+the+Association+for+Computational+Li)
11. Yichen Jiang, Shikha Bordia, Zheng Zhong, Charles Dognin, Maneesh Singh, and Mohit Bansal. 2020. [HoVer: A dataset for many-hop fact extraction and claim verification](https://md.research.elamur.ai/free/en/papers/jiang-2020-hover-dataset-many-hop-fact/). In Findings of the Association for Computational Linguistics: EMNLP 2020, pages 3441–3460, Online. Association for Computational Ling…
12. Yiqiao Jin, Yeon-Chang Lee, Kartik Sharma, Meng Ye, Karan Sikka, Ajay Divakaran, and Srijan Kumar. 2023. Predicting information pathways across online communities. In Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, KDD ’23, page 1044–1056, New York, NY, USA. Ass… [find](https://scholar.google.com/scholar?q=Yiqiao+Jin%2C+Yeon%2DChang+Lee%2C+Kartik+Sharma%2C+Meng+Ye%2C+Karan+Sikka%2C+Ajay+Divakaran%2C+and+Srijan+Kumar%2E+2023%2E+Predicting+information+pathways+across+online+communities%2E+In+Proceedings+of+the+29th+ACM+SIGKD)
13. Amrith Krishna, Sebastian Riedel, and Andreas Vlachos. 2022. ProoFVer: Natural logic theorem proving for fact verification. Transactions of the Association for Computational Linguistics, 10:1013–1030. [find](https://scholar.google.com/scholar?q=Amrith+Krishna%2C+Sebastian+Riedel%2C+and+Andreas+Vlachos%2E+2022%2E+ProoFVer%3A+Natural+logic+theorem+proving+for+fact+verification%2E+Transactions+of+the+Association+for+Computational+Linguistics%2C+10%3A1013%E2%80%931030%2E)
14. Tian Liang, Zhiwei He, Wenxiang Jiao, Xing Wang, Yan Wang, Rui Wang, Yujiu Yang, Shuming Shi, and Zhaopeng Tu. 2024. [Encouraging divergent thinking in large language models through multi-agent debate](https://md.research.elamur.ai/free/en/papers/liang-2023-encouraging-divergent-thinking-large-language/). In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, pages 17…
15. Zhenghao Liu, Chenyan Xiong, Maosong Sun, and Zhiyuan Liu. 2020. [Fine-grained fact verification with kernel graph attention network](https://md.research.elamur.ai/free/en/papers/liu-2020-fine-grained-fact-verification-kernel/). In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 7342–7351, Online. Association for Computational Linguistics.
16. Jiatong Ma, Linmei Hu, Rang Li, and Wenbo Fu. 2025. Local: Logical and causal fact-checking with llm-based multi-agents. In Proceedings of the ACM on Web Conference 2025, WWW ’25, page 1614–1625, New York, NY, USA. Association for Computing Machinery. [find](https://scholar.google.com/scholar?q=Jiatong+Ma%2C+Linmei+Hu%2C+Rang+Li%2C+and+Wenbo+Fu%2E+2025%2E+Local%3A+Logical+and+causal+fact%2Dchecking+with+llm%2Dbased+multi%2Dagents%2E+In+Proceedings+of+the+ACM+on+Web+Conference+2025%2C+WWW+%E2%80%9925%2C+page+1614%E2%80%931625%2C+New+)
17. Bodhisattwa Prasad Majumder, Sudha Rao, Michel Galley, and Julian McAuley. 2021. Ask what’s missing and what’s useful: Improving clarification question generation using global knowledge. In Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguist… [find](https://scholar.google.com/scholar?q=Bodhisattwa+Prasad+Majumder%2C+Sudha+Rao%2C+Michel+Galley%2C+and+Julian+McAuley%2E+2021%2E+Ask+what%E2%80%99s+missing+and+what%E2%80%99s+useful%3A+Improving+clarification+question+generation+using+global+knowledge%2E+In+Proceeding)
18. Preslav Nakov, David Corney, Maram Hasanain, Firoj Alam, Tamer Elsayed, Alberto Barrón-Cedeño, Paolo Papotti, Shaden Shaar, and Giovanni Da San Martino. 2021. Automated fact-checking for assisting human fact-checkers. In Proceedings of the Thirtieth International Joint Conference on Artificial Intel… [find](https://scholar.google.com/scholar?q=Preslav+Nakov%2C+David+Corney%2C+Maram+Hasanain%2C+Firoj+Alam%2C+Tamer+Elsayed%2C+Alberto+Barr%C3%B3n%2DCede%C3%B1o%2C+Paolo+Papotti%2C+Shaden+Shaar%2C+and+Giovanni+Da+San+Martino%2E+2021%2E+Automated+fact%2Dchecking+for+assisting+hum)
19. Nedjma Ousidhoum, Zhangdie Yuan, and Andreas Vlachos. 2022. Varifocal question generation for fact-checking. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pages 2532–2544, Abu Dhabi, United Arab Emirates. Association for Computational Linguistics. [find](https://scholar.google.com/scholar?q=Nedjma+Ousidhoum%2C+Zhangdie+Yuan%2C+and+Andreas+Vlachos%2E+2022%2E+Varifocal+question+generation+for+fact%2Dchecking%2E+In+Proceedings+of+the+2022+Conference+on+Empirical+Methods+in+Natural+Language+Processing%2C+)
20. Liangming Pan, Xiaobao Wu, Xinyuan Lu, Anh Tuan Luu, William Yang Wang, Min-Yen Kan, and Preslav Nakov. 2023. [Fact-checking complex claims with program-guided reasoning](https://md.research.elamur.ai/free/en/papers/pan-2023-fact-checking-complex-claims-program/). In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 6981–700…
21. Sudha Rao and Hal Daumé III. 2019. Answer-based Adversarial Training for Generating Clarification Questions. In 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), pages 143–… [find](https://scholar.google.com/scholar?q=Sudha+Rao+and+Hal+Daum%C3%A9+III%2E+2019%2E+Answer%2Dbased+Adversarial+Training+for+Generating+Clarification+Questions%2E+In+Proceedings+of+the+2019+Conference+of+the+North+American+Chapter+of+the+Association+for+)
22. Arkadiy Saakyan, Tuhin Chakrabarty, and Smaranda Muresan. 2021. [COVID-fact: Fact extraction and verification of real-world claims on COVID-19 pandemic](https://md.research.elamur.ai/free/en/papers/saakyan-2021-covid-fact-fact-extraction-verification/). In Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural La…
23. Tal Schuster, Adam Fisch, and Regina Barzilay. 2021. [Get your vitamin C! robust fact verification with contrastive evidence](https://md.research.elamur.ai/free/en/papers/schuster-2021-get-your-vitamin-c-robust/). In Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 624–643, Online. Assoc…
24. James Thorne and Andreas Vlachos. 2018. [Automated fact checking: Task formulations, methods and future directions](https://md.research.elamur.ai/free/en/papers/thorne-2018-automated-fact-checking-task-formulations/). In Proceedings of the 27th International Conference on Computational Linguistics, pages 3346–3359, Santa Fe, New Mexico, USA. Association for Computational Linguistics.
25. James Thorne, Andreas Vlachos, Christos Christodoulopoulos, and Arpit Mittal. 2018. [FEVER: a large-scale dataset for fact extraction and VERification](https://md.research.elamur.ai/free/en/papers/thorne-2018-fever-large-scale-dataset-fact/). In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Vo…
26. Lin Tian, Xiuzhen Zhang, and Jey Han Lau. 2023. Metatroll: Few-shot detection of state-sponsored trolls with transformer adapters. In Proceedings of the ACM Web Conference 2023, WWW ’23, page 1743–1753, New York, NY, USA. Association for Computing Machinery. [find](https://scholar.google.com/scholar?q=Lin+Tian%2C+Xiuzhen+Zhang%2C+and+Jey+Han+Lau%2E+2023%2E+Metatroll%3A+Few%2Dshot+detection+of+state%2Dsponsored+trolls+with+transformer+adapters%2E+In+Proceedings+of+the+ACM+Web+Conference+2023%2C+WWW+%E2%80%9923%2C+page+1743%E2%80%93175)
27. Xiaoyan Zhao, Lingzhi Wang, Zhanghao Wang, Hong Cheng, Rui Zhang, and Kam-Fai Wong. 2024. PACAR: Automated fact-checking with planning and customized action reasoning using large language models. In Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resourc… [find](https://scholar.google.com/scholar?q=Xiaoyan+Zhao%2C+Lingzhi+Wang%2C+Zhanghao+Wang%2C+Hong+Cheng%2C+Rui+Zhang%2C+and+Kam%2DFai+Wong%2E+2024%2E+PACAR%3A+Automated+fact%2Dchecking+with+planning+and+customized+action+reasoning+using+large+language+models%2E+In+P)

*Updated 1 November 2026*