FaithDial: A Faithful Benchmark for Information-Seeking Dialogue
Open accessThe DOI matches its Crossref record; 42 of the 42 references it cites were found in a registry or an index.
Abstract
The goal of information-seeking dialogue is to respond to seeker queries with natural language utterances that are grounded on knowledge sources. However, dialogue systems often produce unsupported utterances, a phenomenon known as hallucination. To mitigate this behavior, we adopt a data-centric solution and create FaithDial, a new benchmark for hallucination-free dialogues, by editing hallucinated responses in the Wizard of Wikipedia (WoW) benchmark. We observe that FaithDial is more faithful than WoW while also maintaining engaging conversations. We show that FaithDial can serve as training signal for: i) a hallucination critic, which discriminates whether an utterance is faithful or not, and boosts the performance by 12.8 F1 score on the BEGIN benchmark compared to existing datasets for dialogue coherence; ii) high-quality dialogue generation. We benchmark a series of state-of-the-art models and propose an auxiliary contrastive objective that achieves the highest level of faithfulness and abstractiveness based on several automated metrics. Further, we find that the benefits of FaithDial generalize to zero-shot transfer on other datasets, such as CMU-Dog and TopicalChat. Finally, human evaluation reveals that responses generated by models trained on FaithDial are perceived as more interpretable, cooperative, and engaging.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
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
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
Lewis, 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×
“DoHA augments BART (Lewis et al., 2020) with a two-view attention mechanism that separately handles the knowledge document and the dialogue history during generation.”
BackgroundTransformers: State-of-the-Art Natural Language Processing
Wolf, T., Debut, L., Sanh, V., Chaumond, J., Delangue, C., Moi, A., Cistac, P., Rault, T., Louf, R., Funtowicz, M., et al. (2020) · Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing: System Demonstrations · doi:10.18653/v1/2020.emnlp-demos.6 · Cited by 8,741
On the Dangers of Stochastic Parrots: Can Language Models Be Too Big? 🦜
Bender, E. M., Gebru, T., McMillan-Major, A., & Shmitchell, S. (2021) · Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency · doi:10.1145/3442188.3445922 · Cited by 6,988 · cited 1×
“By mimicking the distributional properties of the data, models are bound to ‘‘parrot’’ the hallucinated signals at test time (Bender et al., 2021).”
BackgroundSurvey of Hallucination in Natural Language Generation
Ji, Z., Lee, N., Frieske, R., Yu, T., Su, D., Xu, Y., Ishii, E., Bang, Y. J., Madotto, A., & Fung, P. (2023) · ACM Computing Surveys · doi:10.1145/3571730 · Cited by 5,025 · cited 1×
“Hallucination in knowledge-grounded neural language generation has recently received increasing attention from the NLP community (Ji et al., 2022).”
BackgroundSurvey of Hallucination in Natural Language Generation
Ji, Z., Lee, N., Frieske, R., Yu, T., Su, D., Xu, Y., Ishii, E., Bang, Y., Chen, D., Dai, W., et al. (2022) · arXiv · doi:10.48550/arxiv.2202.03629 · Cited by 4,981
A Broad-Coverage Challenge Corpus for Sentence Understanding through Inference
Williams, A., Nangia, N., & Bowman, S. (2018) · Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long Papers) · doi:10.18653/v1/n18-1101 · Cited by 3,342 · cited 7×
“Surprisingly, MNLI transfers well to F AITH C RITIC .”
BackgroundMethodInfluentialBERTScore: Evaluating Text Generation with BERT
Zhang, T., Kishore, V., Wu, F., Weinberger, K. Q., & Artzi, Y. (2019) · arXiv · doi:10.48550/arxiv.1904.09675 · Cited by 2,865
Cognitive load theory, educational research, and instructional design: some food for thought
de Jong, T. (2010) · Instructional Science · doi:10.1007/s11251-009-9110-0 · Cited by 1,180 · cited 1×
“Moreover, knowledge snippets in W O W tend to be shorter, which is preferable as longer knowledge is correlated with increased hallucination due to the constrained cognitive capacity for text navigation and comprehension in humans (De Jong, 2010; DeStefano and LeFevre, 2007).”
Background
Research works printed but not linked: 22
- Cao, 2021, CLIFF: Contrastive learning for improving faithfulness and factuality in abstractive summarization, 6633, 10.18653/v1/2021.emnlp-main.532 · DOI · Paper page
- DeStefano, 2007, Cognitive load in hypertext reading: A review, Computers in Human Behavior, 23, 1616, 10.1016/j.chb.2005.08.012 · DOI · Paper page
- Dinan, 2019, Wizard of Wikipedia: Knowledge- powered conversational agents
- Dziri, 2019, Evaluating coherence in dialogue systems using entailment, 3806, 10.18653/v1/N19-1381 · DOI · Paper page
- Grusky, 2018, Newsroom: A dataset of 1.3 million summaries with diverse extractive strategies, 708, 10.18653/v1/N18-1065 · DOI · Paper page
- Kang, 2020, Improved natural language generation via loss truncation, 718, 10.18653/v1/2020.acl-main.66 · DOI · Paper page
- Richard Landis, 1977, The measurement of observer agreement for categorical data, biometrics, 159, 10.2307/2529310 · DOI · Paper page
- Li, 2021, Addressing semantic drift in generative question answering with auxiliary extraction, 942, 10.18653/v1/2021.acl-short.118 · DOI
- Lin, 2004, ROUGE: A package for automatic evaluation of summaries, 74
- Liu, 2019, Roberta: A robustly optimized BERT pretraining approach, CoRR
- Lux, 2020, Truth or error? Towards systematic analysis of factual errors in abstractive summaries, 1, 10.18653/v1/2020.eval4nlp-1.1 · DOI · Paper page
- Nie, 2021, I like fish, especially dolphins: Addressing contradictions in dialogue modeling, 1699, 10.18653/v1/2021.acl-long.134 · DOI · Paper page
- van den Oord, 2018, Representation learning with contrastive predictive coding, CoRR
- Papineni, 2002, BLEU: A method for automatic evaluation of machine translation, 311, 10.3115/1073083.1073135 · DOI · Paper page
- Prabhumoye, 2021, Focused attention improves document-grounded generation, 4274, 10.18653/v1/2021.naacl-main.338 · DOI
- Radford, 2019, Language models are unsupervised multitask learners, OpenAI Blog, 1, 9
- Sipos, 2012, Large-margin learning of submodular summarization models, 224
- Stiles, 1992, Describing Talk: A Taxonomy of Verbal Response Modes
- Welleck, 2019, Dialogue natural language inference, 3731, 10.18653/v1/P19-1363 · DOI · Paper page
- Zhang, 2021, On the inductive bias of masked language modeling: From statistical to syntactic dependencies, 5131, 10.18653/v1/2021.naacl-main.404 · DOI
- Zhao, 2020, Knowledge-grounded dialogue generation with pre-trained language models, 3377, 10.18653/v1/2020.emnlp-main.272 · DOI · Paper page
- Zhou, 2018, A dataset for document grounded conversations, 708, 10.18653/v1/D18-1076 · DOI · Paper page
Cited by
HaluEval: A Large-Scale Hallucination Evaluation Benchmark for Large Language Models
Li, J., Cheng, X., Zhao, W. X., Nie, J.-Y., & Wen, J.-R. (2023) · arXiv · doi:10.48550/arxiv.2305.11747 · Cited by 340
“For instance, The BEGIN benchmark (Dziri et al., 2022b) classifies the utterances generated by dialogue systems into three categories, i.e., fully attributable, not fully attributable, and generic; and the Attributable to Identified Sources (AIS) benchmark (Rashkin et al., 2021) assesses whether…”
BackgroundHaluEval: A Large-Scale Hallucination Evaluation Benchmark for Large Language Models
Li, J., Cheng, X., Zhao, X., Nie, J.-Y., & Wen, J.-R. (2023) · Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing · doi:10.18653/v1/2023.emnlp-main.397 · Cited by 343
Evaluating Correctness and Faithfulness of Instruction-Following Models for Question Answering
Adlakha, V., BehnamGhader, P., Lu, X. H., Meade, N., & Reddy, S. (2024) · Transactions of the Association for Computational Linguistics · doi:10.1162/tacl_a_00667 · Cited by 74
“However, they struggle to be faithful to provided knowledge, often failing to accurately identify its relevance.”
BackgroundInfluentialSelf-citationEvaluating Correctness and Faithfulness of Instruction-Following Models for Question Answering
Adlakha, V., BehnamGhader, P., Lu, X. H., Meade, N., & Reddy, S. (2023) · arXiv · doi:10.48550/arxiv.2307.16877 · Cited by 74
Self-citationEvaluating Attribution in Dialogue Systems: The BEGIN Benchmark
Dziri, N., Rashkin, H., Linzen, T., & Reitter, D. (2021) · arXiv · doi:10.48550/arxiv.2105.00071 · Cited by 47
Self-citationVLATest: Testing and Evaluating Vision-Language-Action Models for Robotic Manipulation
Wang, Z., Zhou, Z., Song, J., Huang, Y., Shu, Z., & Ma, L. (2025) · Proceedings of the ACM on Software Engineering · doi:10.1145/3729343 · Cited by 14
HALoGEN: Fantastic LLM Hallucinations and Where to Find Them
Ravichander, A., Ghela, S., Wadden, D., & Choi, Y. (2025) · Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · doi:10.18653/v1/2025.acl-long.71 · Cited by 13
Decision-Making with Deliberation: Meta-reviewing as a Document-grounded Dialogue
Purkayastha, S., Dycke, N., Lauscher, A., & Gurevych, I. (2026) · Proceedings of the 19th Conference of the European Chapter of the Association for Computational Linguistics (Volume 1: Long Papers) · doi:10.18653/v1/2026.eacl-long.158 · Cited by 2
Deceptive Grounding: Entity Attribution Failure in Clinical Retrieval-Augmented Generation
Caruzzo, C., Yoo, D., & Kim, T. S. (2026) · arXiv · doi:10.48550/arxiv.2607.09349
“…[Maynez et al., 2020, Ji et al., 2023, Huang et al., 2025] and that RAG responses accurately reflect retrieved content with genuine citations [Gao et al., 2023, Asai et al., 2024, Es et al., 2024, Min et al., 2023, Kamoi et al., 2023, Stolfo, 2024, Saad-Falcon et al., 2024, Dziri et al., 2022].”
Elastic Weight Removal for Faithful and Abstractive Dialogue Generation
Daheim, N., Dziri, N., Sachan, M., Gurevych, I., & Ponti, E. M. (2023) · arXiv · doi:10.48550/arxiv.2303.17574
Self-citationDeepSeek-R1 Thoughtology: Let's think about LLM Reasoning
Marjanović, S. V., Patel, A., Adlakha, V., Aghajohari, M., BehnamGhader, P., Bhatia, M., Khandelwal, A., Kraft, A., Krojer, B., Lù, X. H., et al. (2025) · arXiv · doi:10.48550/arxiv.2504.07128
Self-citation
Metadata
- Authors
- Dziri, N., Kamalloo, E., Milton, S., Zaiane, O., Yu, M., Ponti, E. M., & Reddy, S.
- Venue
- Transactions of the Association for Computational Linguistics
- Publisher
- MIT Press
- Published
- 2022
- Volume
- 10
- Pages
- 1473-1490
- ISSN
- 2307-387X (ISSN-L)
- OpenAlex
- W4313459318
- Cited by
- 63
- References
- 42
- FWCI
- 3.96
- Type
- journal article
- Language
- en
- Open access
- diamond
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
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.