---
title: "Increasing Faithfulness in Knowledge-Grounded Dialogue with Controllable Features (2021)"
description: "Rather than collecting cleaner data, the authors define a stricter sub-task: the response must be faithful to a… 29 references, cited by 68 papers here."
authors: ["Hannah Rashkin", "David Reitter", "Gaurav Singh Tomar", "Dipanjan Das"]
year: 2021
venue: "Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers)"
source: "https://aclanthology.org/2021.acl-long.58/"
doi: "10.18653/v1/2021.acl-long.58"
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/rashkin-2021-increasing-faithfulness-knowledge-grounded-dialogue/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/rashkin-2021-increasing-faithfulness-knowledge-grounded-dialogue/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2021](https://md.research.elamur.ai/free/en/papers/2021/page/3/) / Increasing Faithfulness in Knowledge-Grounded Dialogue with Controllable Features

# Increasing Faithfulness in Knowledge-Grounded Dialogue with Controllable Features

**Authors:** [Hannah Rashkin](https://md.research.elamur.ai/free/en/authors/hannah-rashkin/), [David Reitter](https://md.research.elamur.ai/free/en/authors/david-reitter/), [Gaurav Singh Tomar](https://md.research.elamur.ai/free/en/authors/gaurav-singh-tomar/) and [Dipanjan Das](https://md.research.elamur.ai/free/en/authors/dipanjan-das/) · **Year:** 2021 · **Venue:** [Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers)](https://md.research.elamur.ai/free/en/venues/acl/) ([ACL 2021](https://md.research.elamur.ai/free/en/venues/acl/2021/)) · **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/2021.acl-long.58/) · [PDF](https://aclanthology.org/2021.acl-long.58.pdf) · [DOI 10.18653/v1/2021.acl-long.58](https://doi.org/10.18653/v1/2021.acl-long.58) · [OpenAlex](https://openalex.org/W3176098057)

**Citations:** 72 citations (OpenAlex, 2026-10-14) · [68](https://md.research.elamur.ai/free/en/papers/rashkin-2021-increasing-faithfulness-knowledge-grounded-dialogue/cited-by/) in our library, of the 141,840 papers whose bibliographies we read

## Summary
Knowledge-grounded dialogue agents are meant to pass on information from a trusted source text. Datasets such as Wizard of Wikipedia, however, mix evidence-based replies with chit-chat, opinions and invented personal details, so models trained on them tend to hallucinate. Rather than collecting cleaner data, the authors define a stricter sub-task: the response must be faithful to a given gold evidence span and still fit the conversation.

They score responses with three measures:
- objective voice (no first-person singular pronouns),
- unigram lexical precision against the evidence,
- entailment predicted by a RoBERTa model trained on MNLI.

During fine-tuning of GPT-2 and T5, these scores are added to the input as control-code tokens computed from the gold reply. At inference, the codes are fixed to their most grounded values. A second technique resamples outputs at decoding time until they meet the constraints.

Human raters judged the controlled models more objective and more faithful to the evidence than the baselines. The proposed measures also correlated with these human judgements. This shows that noisy grounded corpora can still yield trustworthy, informative dialogue agents.

## Key points
- Data: Wizard of Wikipedia, using only wizard replies to the apprentice: 73,571 training responses. Dev has 3,905 seen-topic and 3,898 unseen-topic responses; test has 3,842 seen and 3,902 unseen.
- Training-set analysis: 44% of responses use first person, average lexical precision against the evidence is 0.43, and only 23% are predicted as entailed. Simply filtering the data would therefore discard most of it.
- Control tokens such as `<high-prec>`, `<no-1st-person>` and `<entail>` are prepended to the evidence and conversation-history sequence, in the style of CTRL.
- NLI neutral and contradiction labels are merged into a single "non-entailing" class.
- Evaluation combines BLEU, the three proposed measures, and a human study focused on faithfulness and objectivity.

*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

- [Google](https://md.research.elamur.ai/free/en/companies/google/) — Company, Mountain View, United States · 353 works · 364,806 citations

📚 Shelves (4): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 9,086 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 8,889 works · [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 9,296 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 9,206 works

🧭 Topics (1): [🗨️ Dialogue systems](https://md.research.elamur.ai/free/en/topics/dialogue-systems/) · 401 works

## Cite

Cite as: Hannah Rashkin, David Reitter, Gaurav Singh Tomar and Dipanjan Das. 2021. Increasing Faithfulness in Knowledge-Grounded Dialogue with Controllable Features. Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers).

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

```bibtex
@inproceedings{rashkin2021increasing,
  title = {{Increasing Faithfulness in Knowledge-Grounded Dialogue with Controllable Features}},
  author = {Hannah Rashkin, David Reitter, Gaurav Singh Tomar and Dipanjan Das},
  booktitle = {Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers)},
  year = {2021},
  doi = {10.18653/v1/2021.acl-long.58},
  url = {https://doi.org/10.18653/v1/2021.acl-long.58}
}
```

- [← previous SemEval 2021 Task 7: HaHackathon, Detecting and Rating Humor and Offense](https://md.research.elamur.ai/free/en/papers/meaney-2021-semeval-2021-task-7-hahackathon/)
- [next → Paragraph-level Simplification of Medical Texts](https://md.research.elamur.ai/free/en/papers/devaraj-2021-paragraph-level-simplification-medical-texts/)
- [↑ back to papers of 2021](https://md.research.elamur.ai/free/en/papers/2021/page/3/)

## References

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

1. Ziqiang Cao, Furu Wei, Wenjie Li, and Sujian Li. 2018. Faithful to the original: Fact aware neural abstractive summarization. In Proceedings of the Thirty-Second AAAI Conference on Artificial Intelligence, (AAAI-18), the 30th innovative Applications of Artificial Intelligence (IAAI-18), and the 8th… [find](https://scholar.google.com/scholar?q=Ziqiang+Cao%2C+Furu+Wei%2C+Wenjie+Li%2C+and+Sujian+Li%2E+2018%2E+Faithful+to+the+original%3A+Fact+aware+neural+abstractive+summarization%2E+In+Proceedings+of+the+Thirty%2DSecond+AAAI+Conference+on+Artificial+Intellig)
2. Emily Dinan, Stephen Roller, Kurt Shuster, Angela Fan, Michael Auli, and Jason Weston. 2019. Wizard of wikipedia: Knowledge-powered conversational agents. In 7th International Conference on Learning Representations, ICLR 2019, New Orleans, LA, USA, May 6-9, 2019. OpenReview.net. [find](https://scholar.google.com/scholar?q=Emily+Dinan%2C+Stephen+Roller%2C+Kurt+Shuster%2C+Angela+Fan%2C+Michael+Auli%2C+and+Jason+Weston%2E+2019%2E+Wizard+of+wikipedia%3A+Knowledge%2Dpowered+conversational+agents%2E+In+7th+International+Conference+on+Learning+R)
3. Tobias Falke, Leonardo F. R. 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 Lingui… [our card](https://md.research.elamur.ai/free/en/papers/falke-2019-ranking-generated-summaries-correctness-interesting/)
4. Katja Filippova. 2020. Controlled hallucinations: Learning to generate faithfully from noisy data. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing: Findings, EMNLP 2020, Online Event, 16-20 November 2020, pages 864–870. Association for Computational Linguist… [find](https://scholar.google.com/scholar?q=Katja+Filippova%2E+2020%2E+Controlled+hallucinations%3A+Learning+to+generate+faithfully+from+noisy+data%2E+In+Proceedings+of+the+2020+Conference+on+Empirical+Methods+in+Natural+Language+Processing%3A+Findings%2C+)
5. Marjan Ghazvininejad, Chris Brockett, Ming-Wei Chang, Bill Dolan, Jianfeng Gao, Wen-tau Yih, and Michel Galley. 2018. A knowledge-grounded neural conversation model. In Proceedings of the Thirty-Second AAAI Conference on Artificial Intelligence, (AAAI-18), the 30th innovative Applications of Artific… [find](https://scholar.google.com/scholar?q=Marjan+Ghazvininejad%2C+Chris+Brockett%2C+Ming%2DWei+Chang%2C+Bill+Dolan%2C+Jianfeng+Gao%2C+Wen%2Dtau+Yih%2C+and+Michel+Galley%2E+2018%2E+A+knowledge%2Dgrounded+neural+conversation+model%2E+In+Proceedings+of+the+Thirty%2DSecon)
6. Karthik Gopalakrishnan, Behnam Hedayatnia, Qinlang Chen, Anna Gottardi, Sanjeev Kwatra, Anu Venkatesh, Raefer Gabriel, and Dilek Hakkani-Tür. 2019. [find](https://scholar.google.com/scholar?q=Karthik+Gopalakrishnan%2C+Behnam+Hedayatnia%2C+Qinlang+Chen%2C+Anna+Gottardi%2C+Sanjeev+Kwatra%2C+Anu+Venkatesh%2C+Raefer+Gabriel%2C+and+Dilek+Hakkani%2DTu%CC%88r%2E+2019%2E)
7. Topical-Chat: Towards Knowledge-Grounded Open-Domain Conversations. In Proc. Interspeech 2019, pages 1891–1895. [find](https://scholar.google.com/scholar?q=Topical%2DChat%3A+Towards+Knowledge%2DGrounded+Open%2DDomain+Conversations%2E+In+Proc%2E+Interspeech+2019%2C+pages+1891%E2%80%931895%2E)
8. Ari Holtzman, Jan Buys, Li Du, Maxwell Forbes, and Yejin Choi. 2020. The curious case of neural text degeneration. In 8th International Conference on Learning Representations, ICLR 2020, Addis Ababa, Ethiopia, April 26-30, 2020. OpenReview.net. [find](https://scholar.google.com/scholar?q=Ari+Holtzman%2C+Jan+Buys%2C+Li+Du%2C+Maxwell+Forbes%2C+and+Yejin+Choi%2E+2020%2E+The+curious+case+of+neural+text+degeneration%2E+In+8th+International+Conference+on+Learning+Representations%2C+ICLR+2020%2C+Addis+Ababa%2C+)
9. Nitish Shirish Keskar, Bryan McCann, Lav R. Varshney, Caiming Xiong, and Richard Socher. 2019. CTRL: A conditional transformer language model for controllable generation. CoRR, abs/1909.05858. [our card](https://md.research.elamur.ai/free/en/papers/keskar-2019-ctrl-conditional-transformer-language-model/)
10. Byeongchang Kim, Jaewoo Ahn, and Gunhee Kim. 2020. Sequential latent knowledge selection for knowledge-grounded dialogue. In 8th International Conference on Learning Representations, ICLR 2020, Addis Ababa, Ethiopia, April 26-30, 2020. OpenReview.net. [find](https://scholar.google.com/scholar?q=Byeongchang+Kim%2C+Jaewoo+Ahn%2C+and+Gunhee+Kim%2E+2020%2E+Sequential+latent+knowledge+selection+for+knowledge%2Dgrounded+dialogue%2E+In+8th+International+Conference+on+Learning+Representations%2C+ICLR+2020%2C+Addis+)
11. Rongzhong Lian, Min Xie, Fan Wang, Jinhua Peng, and Hua Wu. 2019. Learning to select knowledge for response generation in dialog systems. In Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence, IJCAI 2019, Macao, China, August 10-16, 2019, pages 5081– 5087. [find](https://scholar.google.com/scholar?q=Rongzhong+Lian%2C+Min+Xie%2C+Fan+Wang%2C+Jinhua+Peng%2C+and+Hua+Wu%2E+2019%2E+Learning+to+select+knowledge+for+response+generation+in+dialog+systems%2E+In+Proceedings+of+the+Twenty%2DEighth+International+Joint+Confer)
12. Chia-Wei Liu, Ryan Lowe, Iulian Serban, Mike Noseworthy, Laurent Charlin, and Joelle Pineau. 2016. How NOT to evaluate your dialogue system: An empirical study of unsupervised evaluation metrics for dialogue response generation. In Proceedings of the 2016 Conference on Empirical Methods in Natural L… [our card](https://md.research.elamur.ai/free/en/papers/liu-2016-how-not-evaluate-your-dialogue/)
13. Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, and Veselin Stoyanov. 2019. Roberta: A robustly optimized BERT pretraining approach. CoRR, abs/1907.11692. [find](https://scholar.google.com/scholar?q=Yinhan+Liu%2C+Myle+Ott%2C+Naman+Goyal%2C+Jingfei+Du%2C+Mandar+Joshi%2C+Danqi+Chen%2C+Omer+Levy%2C+Mike+Lewis%2C+Luke+Zettlemoyer%2C+and+Veselin+Stoyanov%2E+2019%2E+Roberta%3A+A+robustly+optimized+BERT+pretraining+approach%2E+C)
14. Joshua Maynez, Shashi Narayan, Bernd Bohnet, and Ryan McDonald. 2020. On faithfulness and factuality in abstractive summarization. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 1906–1919, Online. Association for Computational Linguistics. [find](https://scholar.google.com/scholar?q=Joshua+Maynez%2C+Shashi+Narayan%2C+Bernd+Bohnet%2C+and+Ryan+McDonald%2E+2020%2E+On+faithfulness+and+factuality+in+abstractive+summarization%2E+In+Proceedings+of+the+58th+Annual+Meeting+of+the+Association+for+Comp)
15. Chuan Meng, Pengjie Ren, Zhumin Chen, Weiwei Sun, Zhaochun Ren, Zhaopeng Tu, and Maarten de Rijke. 2020. Dukenet: A dual knowledge interaction network for knowledge-grounded conversation. In Proceedings of the 43rd International ACM SIGIR conference on research and development in Information Retriev… [find](https://scholar.google.com/scholar?q=Chuan+Meng%2C+Pengjie+Ren%2C+Zhumin+Chen%2C+Weiwei+Sun%2C+Zhaochun+Ren%2C+Zhaopeng+Tu%2C+and+Maarten+de+Rijke%2E+2020%2E+Dukenet%3A+A+dual+knowledge+interaction+network+for+knowledge%2Dgrounded+conversation%2E+In+Proceedin)
16. Nikita Moghe, Siddhartha Arora, Suman Banerjee, and Mitesh M. Khapra. 2018. Towards exploiting background knowledge for building conversation systems. In Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing, pages 2322–2332, Brussels, Belgium. Association for Comput… [our card](https://md.research.elamur.ai/free/en/papers/moghe-2018-exploiting-background-knowledge-building-conversation/)
17. Tong Niu and Mohit Bansal. 2018. Polite dialogue generation without parallel data. Transactions of the Association for Computational Linguistics, 6:373–389. [find](https://scholar.google.com/scholar?q=Tong+Niu+and+Mohit+Bansal%2E+2018%2E+Polite+dialogue+generation+without+parallel+data%2E+Transactions+of+the+Association+for+Computational+Linguistics%2C+6%3A373%E2%80%93389%2E)
18. Ratish Puduppully, Li Dong, and Mirella Lapata. 2019. Data-to-text generation with content selection and planning. Proceedings of the AAAI Conference on Artificial Intelligence, 33(01):6908–6915. [find](https://scholar.google.com/scholar?q=Ratish+Puduppully%2C+Li+Dong%2C+and+Mirella+Lapata%2E+2019%2E+Data%2Dto%2Dtext+generation+with+content+selection+and+planning%2E+Proceedings+of+the+AAAI+Conference+on+Artificial+Intelligence%2C+33%2801%29%3A6908%E2%80%936915%2E)
19. Lianhui Qin, Michel Galley, Chris Brockett, Xiaodong Liu, Xiang Gao, Bill Dolan, Yejin Choi, and Jianfeng Gao. 2019. Conversing by reading: Contentful neural conversation with on-demand machine reading. In ACL, pages 5427–5436, Florence, Italy. Association for Computational Linguistics. [find](https://scholar.google.com/scholar?q=Lianhui+Qin%2C+Michel+Galley%2C+Chris+Brockett%2C+Xiaodong+Liu%2C+Xiang+Gao%2C+Bill+Dolan%2C+Yejin+Choi%2C+and+Jianfeng+Gao%2E+2019%2E+Conversing+by+reading%3A+Contentful+neural+conversation+with+on%2Ddemand+machine+readin)
20. Alec Radford, Jeff Wu, Rewon Child, David Luan, Dario Amodei, and Ilya Sutskever. 2019. Language models are unsupervised multitask learners. [find](https://scholar.google.com/scholar?q=Alec+Radford%2C+Jeff+Wu%2C+Rewon+Child%2C+David+Luan%2C+Dario+Amodei%2C+and+Ilya+Sutskever%2E+2019%2E+Language+models+are+unsupervised+multitask+learners%2E)
21. Colin Raffel, Noam Shazeer, Adam Roberts, Katherine Lee, Sharan Narang, Michael Matena, Yanqi Zhou, Wei Li, and Peter J. Liu. 2020. Exploring the limits of transfer learning with a unified text-to-text transformer. J. Mach. Learn. Res., 21:140:1–140:67. [our card](https://md.research.elamur.ai/free/en/papers/raffel-2019-exploring-limits-transfer-learning-unified/)
22. Abigail See, Stephen Roller, Douwe Kiela, and Jason Weston. 2019. What makes a good conversation? how controllable attributes affect human judgments. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Vol… [our card](https://md.research.elamur.ai/free/en/papers/see-2019-what-makes-good-conversation-how/)
23. Kurt Shuster, Da Ju, Stephen Roller, Emily Dinan, Y-Lan Boureau, and Jason Weston. 2020. The dialogue dodecathlon: Open-domain knowledge and image grounded conversational agents. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 2453–2470, Online. Asso… [find](https://scholar.google.com/scholar?q=Kurt+Shuster%2C+Da+Ju%2C+Stephen+Roller%2C+Emily+Dinan%2C+Y%2DLan+Boureau%2C+and+Jason+Weston%2E+2020%2E+The+dialogue+dodecathlon%3A+Open%2Ddomain+knowledge+and+image+grounded+conversational+agents%2E+In+Proceedings+of+the)
24. Zhiliang Tian, Wei Bi, Dongkyu Lee, Lanqing Xue, Yiping Song, Xiaojiang Liu, and Nevin L. Zhang. 2020. Response-anticipated memory for on-demand knowledge integration in response generation. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, ACL 2020, Online,… [find](https://scholar.google.com/scholar?q=Zhiliang+Tian%2C+Wei+Bi%2C+Dongkyu+Lee%2C+Lanqing+Xue%2C+Yiping+Song%2C+Xiaojiang+Liu%2C+and+Nevin+L%2E+Zhang%2E+2020%2E+Response%2Danticipated+memory+for+on%2Ddemand+knowledge+integration+in+response+generation%2E+In+Procee)
25. Thomas Wolf, Lysandre Debut, Victor Sanh, Julien Chaumond, Clement Delangue, Anthony Moi, Pier-ric Cistac, Tim Rault, Rémi Louf, Morgan Funtowicz, Joe Davison, Sam Shleifer, Patrick von Platen, Clara Ma, Yacine Jernite, Julien Plu, Canwen Xu, Teven Le Scao, Sylvain Gugger, Mariama Drame, Quentin Lh… [our card](https://md.research.elamur.ai/free/en/papers/wolf-2020-transformers-state-art-natural-language/)
26. Zeqiu Wu, Michel Galley, Chris Brockett, Yizhe Zhang, Xiang Gao, Chris Quirk, Rik Koncel-Kedziorski, Jianfeng Gao, Hannaneh Hajishirzi, Mari Ostendorf, and Bill Dolan. 2020. A controllable model of grounded response generation. CoRR, abs/2005.00613. — [arXiv:2005.00613](https://arxiv.org/abs/2005.00613) [find](https://scholar.google.com/scholar?q=Zeqiu+Wu%2C+Michel+Galley%2C+Chris+Brockett%2C+Yizhe+Zhang%2C+Xiang+Gao%2C+Chris+Quirk%2C+Rik+Koncel%2DKedziorski%2C+Jianfeng+Gao%2C+Hannaneh+Hajishirzi%2C+Mari+Ostendorf%2C+and+Bill+Dolan%2E+2020%2E+A+controllable+model+of+gr)
27. Saizheng Zhang, Emily Dinan, Jack Urbanek, Arthur Szlam, Douwe Kiela, and Jason Weston. 2018. Personalizing dialogue agents: I have a dog, do you have pets too? In Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 2204– 2213, Melbo… [our card](https://md.research.elamur.ai/free/en/papers/zhang-2018-personalizing-dialogue-agents-i-have/)
28. Yizhe Zhang, Siqi Sun, Michel Galley, Yen-Chun Chen, Chris Brockett, Xiang Gao, Jianfeng Gao, Jingjing Liu, and Bill Dolan. 2020. DIALOGPT : Large-scale generative pre-training for conversational response generation. In Proceedings of the 58th Annual Meeting of the Association for Computational Ling… [our card](https://md.research.elamur.ai/free/en/papers/zhang-2019-dialogpt-large-scale-generative-pre/)
29. Zheng Zhao, Shay B. Cohen, and Bonnie Webber. 2020. Reducing quantity hallucinations in abstractive summarization. In Findings of the Association for Computational Linguistics: EMNLP 2020, pages 2237– 2249, Online. Association for Computational Linguistics. [our card](https://md.research.elamur.ai/free/en/papers/zhao-2020-reducing-quantity-hallucinations-abstractive-summarization/)

*Updated 24 October 2026*