SummEval: Re-evaluating Summarization Evaluation
Open accessThe DOI matches its Crossref record; 99 of the 102 references it cites were found in a registry or an index.
Abstract
The scarcity of comprehensive up-to-date studies on evaluation metrics for text summarization and the lack of consensus regarding evaluation protocols continue to inhibit progress. We address the existing shortcomings of summarization evaluation methods along five dimensions: 1) we re-evaluate 14 automatic evaluation metrics in a comprehensive and consistent fashion using neural summarization model outputs along with expert and crowd-sourced human annotations; 2) we consistently benchmark 23 recent summarization models using the aforementioned automatic evaluation metrics; 3) we assemble the largest collection of summaries generated by models trained on the CNN/DailyMail news dataset and share it in a unified format; 4) we implement and share a toolkit that provides an extensible and unified API for evaluating summarization models across a broad range of automatic metrics; and 5) we assemble and share the largest and most diverse, in terms of model types, collection of human judgments of model-generated summaries on the CNN/Daily Mail dataset annotated by both expert judges and crowd-source workers. We hope that this work will help promote a more complete evaluation protocol for text summarization as well as advance research in developing evaluation metrics that better correlate with human judgments.
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.
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
“SummEval (Fabbri et al., 2021) is a news domain dataset, where the task is abstractive summarization.The outputs are systemgenerated summaries of CNN/Daily Mail articles, with human ratings provided on coherence, consistency, fluency, and relevance.”
From Learning to Judge: LLMs Designing and Applying Evaluation Rubrics (2026) · cites this paper as Fabbri et al., 2021
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
“Human evaluation has long been considered the gold standard for evaluating the quality of natural language generation (NLG) systems (Belz and Reiter, 2006;Lai and Tetreault, 2018;Fabbri et al., 2021).”
From SkillAggregation: Reference-free LLM-Dependent Aggregation (2025) · cites this paper as Fabbri et al., 2021
Experimental — accuracy not yet measuredThe cited source says what the claim attributes to it.
In the citing paper
“Indirectly, these metrics also evaluate composition, since it has been shown that ROUGE and BLEU correlate with human judgments on fluency of a text [67].”
From From Anecdotal Evidence to Quantitative Evaluation Methods: A Systematic Review on Evaluating Explainable AI (2023) · cites this paper as [67]
The cited source says what the claim attributes to it.
In the citing paper
“Data in AG-GREFACT comes from 9 factual consistency evaluation datasets on CNN or XSum, including widely used ones such as SummaC (Laban et al., 2022), FRANK (Pagnoni et al., 2021), and SummEval (Fabbri et al., 2021).”
From MiniCheck: Efficient Fact-Checking of LLMs on Grounding Documents (2024) · cites this paper as Fabbri et al., 2021
The cited source says what the claim attributes to it.
In the citing paper
“most prior studies have primarily concentrated on task-specific hallucinations, such as abstractive summarization [81, 102, 152, 208, 236, 310], data-to-text[240, 296], and machine translation [389].”
From A Survey on Hallucination in Large Language Models: Principles, Taxonomy, Challenges, and Open Questions (2025) · cites this paper as [81]
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
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 · cited 1×
“∗Equal contributions from authors The field has benefited from advances in neural network architectures (Sutskever et al., 2014; Bahdanau et al., 2014; Vinyals et al., 2015; Vaswani et al., 2017) as well as the availability of large-scale datasets (Sandhaus, 2008; Hermann et al.”
BackgroundBERT: 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 7×
“Recent advances in pretrained language models, such as BERT (Devlin et al., 2019), have motivated a corresponding shift to pretraining methods in summarization (Liu and Lapata, 2019; Zhang et al.”
BackgroundMethodInfluentialDistributed Representations of Words and Phrases and their Compositionality
Mikolov, T., Sutskever, I., Chen, K., Corrado, G., & Dean, J. (2013) · arXiv · doi:10.48550/arxiv.1310.4546 · Cited by 24,533 · cited 1×
“ROUGE-WE (Ng and Abrecht, 2015) extends ROUGE by using soft lexical matching based on the cosine similarity of Word2Vec (Mikolov et al., 2013) embeddings.”
MethodBLEU: a method for automatic evaluation of machine translation
Papineni, K., Roukos, S., Ward, T., & Zhu, W.-J. (2001) · Proceedings of the 40th Annual Meeting on Association for Computational Linguistics - ACL '02 · doi:10.3115/1073083.1073135 · Cited by 21,344 · cited 3×
“However, it has recently been shown that the mentioned datasets contain human judgments for model outputs scoring on a lower scale compared to current summarization systems putting into question the true performance of those metrics in the new setting (Peyrard, 2019).”
BackgroundInfluentialSequence 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 · cited 1×
“The field has benefited from advances in neural network architectures (Sutskever et al., 2014; Bahdanau et al., 2014; Vinyals et al., 2015; Vaswani ∗ Equal contributions from authors et al., 2017) as well as the availability of largescale datasets (Sandhaus, 2008; Hermann et al., 2015; Grusky et…”
BackgroundProceedings of IEEE Conference on Computer Vision and Pattern Recognition
(1993) · doi:10.1109/cvpr.1993 · Cited by 17,667
Neural Machine Translation by Jointly Learning to Align and Translate
Bahdanau, D., Cho, K., & Bengio, Y. (2014) · arXiv · doi:10.48550/arxiv.1409.0473 · Cited by 16,213 · cited 1×
“∗Equal contributions from authors The field has benefited from advances in neural network architectures (Sutskever et al., 2014; Bahdanau et al., 2014; Vinyals et al., 2015; Vaswani et al., 2017) as well as the availability of large-scale datasets (Sandhaus, 2008; Hermann et al.”
BackgroundAdvances in Neural Information Processing Systems 19
(2007) · The MIT Press eBooks · doi:10.52202/085713 · Cited by 15,539
Advances in Neural Information Processing Systems 29
Teymur, O., Zygalakis, K., & Calderhead, B. (2016) · Cited by 13,887
Advances in Neural Information Processing Systems 25
Zhang, Y., Sutton, C. A., Storkey, A., & Zoubin Ghahramani, A. (2012) · Cited by 12,653
Research works printed but not linked: 2
- Dang, 2009, Overview of the TAC 2009 summarization track
- Radford, 2019, Language models are unsupervised multitask learners, OpenAI Blog, 1, 9
Cited by
Items 301–400 of 762
Guo, Y., Sohn, J. H., Leroy, G., & Cohen, T. (2026) · Journal of Biomedical Informatics · doi:10.1016/j.jbi.2026.105038 · Cited by 9
Kelsall, J., Tan, X., Bergin, A., Chen, J., Waheed, M., Sorell, T., Procter, R., Liakata, M., Chim, J., & Chi, S. (2025) · AI & Society · doi:10.1007/s00146-025-02741-9 · Cited by 16
“Moreover, semantics is only one of the possible dimensions with factuality, evidence appropriateness, and coherence being other aspects to evaluate (Fabbri et al. 2021; Gehrmann et al. 2023; Song et al. 2024).”
Evaluating the quality of GenAI applications in software engineering: a multi-case study
Yu, L., Alégroth, E., Chatzipetrou, P., & Gorschek, T. (2025) · Empirical Software Engineering · doi:10.1007/s10664-025-10759-2 · Cited by 11
Large language models in healthcare and biomedical informatics: A comprehensive review
Hornback, A. M., Sathu, H., Kim, K., Wang, Y., Zhu, Y., Isgut, M., Avula, P., Khimani, A., & Wang, M. D. (2026) · Innovation and Emerging Technologies · doi:10.1142/s2737599426300011 · Cited by 9
Self-Consistency Falls Short! The Adverse Effects of Positional Bias on Long-Context Problems
Byerly, A., & Khashabi, D. (2026) · Transactions of the Association for Computational Linguistics · doi:10.1162/tacl.a.625 · Cited by 12
“While ROUGE quantifies surface-level alignment, it fails to capture essential quality dimensions, such as factual consistency and narrative flow (Fabbri et al., 2021).”
LLM Essay Scoring Under Holistic and Analytic Rubrics: Prompt Effects and Bias
Kucia, F. J., Chakraborty, A., & Wróblewska, A. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-29918-5_38 · Cited by 12
“This approach has been explored across a range of generation tasks, including machine translation, dialogue, and summarization [9,4].”
Multi-Sourced, Multi-Agent Evidence Retrieval for Fact-Checking
Gong, S., Sinnott, R., Qi, J., Paris, C. L., Nakov, P., & Xie, Z. (2026) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3805712.3809686 · Cited by 11
“…retrieval over both structured and un-structured knowledge , by dynamically integrating knowledge • Achieves self-improvement through reflection-driven prompt adaptation , by allowing the agent to refine its policies, WKGFC advances towards robust, open-world, and self-adaptive fact-checking .”
Implementing AI Vertical use cases ‐ Scenario 1
Cagnoni, S., Emiliani, V., Lombardo, G., Alkema, W. B. L., Hooijmans, C. R., Alkema, S. G., Novotny, T., Hair, K., Nič, M., Macleod, M. R., et al. (2023) · EFSA Supporting Publications · doi:10.2903/sp.efsa.2023.en-8223 · Cited by 11
Garg, P. K., Chakraborty, R., & Dandapat, S. K. (2025) · Knowledge-Based Systems · doi:10.1016/j.knosys.2025.112969 · Cited by 9
CriSPO: Multi-Aspect Critique-Suggestion-guided Automatic Prompt Optimization for Text Generation
He, H., Liu, Q., Xu, L., Shivade, C., Zhang, Y., Srinivasan, S., & Kirchhoff, K. (2025) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v39i22.34575 · Cited by 28
MDCure: A Scalable Pipeline for Multi-Document Instruction-Following
Liu, G., Shi, B., Caciularu, A., Szpektor, I., & Cohan, A. (2025) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/2025.acl-long.1418 · Cited by 15
Is Human Scoring the Best Criteria for Summary Evaluation?
Vasilyev, O., & Bohannon, J. (2021) · Findings of the Association for Computational Linguistics: ACL (Findings of ACL) · doi:10.18653/v1/2021.findings-acl.192 · Cited by 10
Chen, Y., Zhang, H., Zhou, Y., Bai, X., Wang, Y., Ming, Z., Yan, J., Li, Y., Li, J., Xianchao, Z., et al. (2023) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/2023.acl-long.519 · Cited by 15
Socratic Pretraining: Question-Driven Pretraining for Controllable Summarization
Pagnoni, A., Fabbri, A. R., Kryściński, W., & Wu, C.-S. (2023) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/2023.acl-long.713 · Cited by 27
Self-citationIn-sample Curriculum Learning by Sequence Completion for Natural Language Generation
Jia, Q. F., Liu, Y., Tang, H., & Zhu, K. Q. (2023) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/2023.acl-long.666 · Cited by 8
DecipherPref: Analyzing Influential Factors in Human Preference Judgments via GPT-4
Hu, Y., Song, K., Cho, S., Wang, X., Foroosh, H., & Liu, F. (2023) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/2023.emnlp-main.519 · Cited by 24
“Evaluators are asked to assess the quality of summaries along certain linguistic and content dimensions; they can also rank outputs from various systems (Bhandari et al., 2020; Fabbri et al., 2021; Clark et al., 2021).”
BackgroundSummary Explorer: Visualizing the State of the Art in Text Summarization
Syed, S., Yousef, T., Al‐Khatib, K., Jänicke, S., & Potthast, M. (2021) · Conference on Empirical Methods in Natural Language Processing: System Demonstrations (EMNLP) · doi:10.18653/v1/2021.emnlp-demo.22 · Cited by 10
What Makes a Good Podcast Summary?
Rezapour, R., Reddy, S., Jones, R., & Soboroff, I. M. (2022) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3477495.3531802 · Cited by 8
“[6] benchmark a suite of models for the CNN/Daily Mail corpus [10], and compare themwith respect to different evaluation metrics and human judgments.”
BackgroundMethodResultShortcomings of Question Answering Based Factuality Frameworks for Error Localization
Kamoi, R., Goyal, T., & Durrett, G. (2023) · Conference of the European Chapter of the Association for Computational Linguistics · doi:10.18653/v1/2023.eacl-main.11 · Cited by 17
Efficient Benchmarking (of Language Models)
Perlitz, Y., Bandel, E., Gera, A., Arviv, O., Ein-Dor, L., Shnarch, E., Slonim, N., Shmueli-Scheuer, M., & Choshen, L. (2024) · Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers) · doi:10.18653/v1/2024.naacl-long.139 · Cited by 7
Comparison-based Active Preference Learning for Multi-dimensional Personalization
Oh, M., Lee, S., & Ok, J. (2025) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/2025.acl-long.1590 · Cited by 11
“We conduct experiments on the following language generation tasks: (1) Assistant on HH-RLHF (Bai et al., 2022a) dataset, (2) Summary on Summarize-from-Feedback (Stiennon et al., 2020) dataset, and (3) Summary+ on SummEval (Fabbri et al., 2020) dataset.”
AgentReport: A Multi-Agent LLM Approach for Automated and Reproducible Bug Report Generation
Choi, S., & Yang, G. (2025) · Applied Sciences · doi:10.3390/app152211931 · Cited by 9
Learning to Rank Utterances for Query-Focused Meeting Summarization
Liu, X., & Xu, Y. (2023) · Findings of the Association for Computational Linguistics: ACL (Findings of ACL) · doi:10.18653/v1/2023.findings-acl.538 · Cited by 7
Annotating and Detecting Fine-grained Factual Errors for Dialogue Summarization
Zhu, R., Qi, J., & Lau, J. H. (2023) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/2023.acl-long.377 · Cited by 18
SumBlogger: Abstractive Summarization of Large Collections of Scientific Articles
Zakkas, P., Verberne, S., & Zavrel, J. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-56027-9_23 · Cited by 16
FEEL: A Framework for Evaluating Emotional Support Capability with Large Language Models
Zhang, H., Chen, Y., Wang, M., & Feng, S. (2024) · Lecture notes in computer science · doi:10.1007/978-981-97-5618-6_9 · Cited by 11
“…by computing probabilities linked to the assignment of particular scoring bands to get stable evaluation scores; and c) a comprehensive integrative weighted computational approach using ensemble learning method that synthesizes results from multiple LLMs to deduce definitive evaluation scores.”
MethodSATS: simplification aware text summarization of scientific documents
Zaman, F., Kamiran, F., Shardlow, M., Hassan, S., Karim, A., & Aljohani, N. R. (2024) · Frontiers in Artificial Intelligence · doi:10.3389/frai.2024.1375419 · Cited by 8
Evaluation of Text Cluster Naming with Generative Large Language Models
Preiss, A., Arbeit, C. A., Berghammer, A., Bollenbacher, J., McCarthy, J., Brom, M. G., Enger, M., Villacorta, N. R., & Straughn, S. (2024) · Journal of Data Science · doi:10.6339/24-jds1149 · Cited by 7
End-to-End Long Document Summarization using Gradient Caching
Saxena, R., Tang, H., & Keller, F. (2025) · Transactions of the Association for Computational Linguistics · doi:10.1162/tacl.a.40 · Cited by 12
NonFactS: NonFactual Summary Generation for Factuality Evaluation in Document Summarization
Soleimani, A., Monz, C., & Worring, M. (2023) · Findings of the Association for Computational Linguistics: ACL (Findings of ACL) · doi:10.18653/v1/2023.findings-acl.400 · Cited by 15
“The SUMMAC benchmark comprises the six largest datasets standardized for factuality evaluation: CGS (Falke et al., 2019), XSF (Maynez et al., 2020), Polytope (Huang et al., 2020), FactCC (Kryscinski et al., 2020), SummEval (Fabbri et al., 2021), and FRANK (Pagnoni et al., 2021).”
MethodFrom task to evaluation: an automatic text summarization review
Lu, L., Liu, Y., Xu, W., Li, H., & Sun, G. (2023) · Artificial Intelligence Review · doi:10.1007/s10462-023-10582-5 · Cited by 10
“Furthermore, Fabbri et al. (2021) re-evaluated the efficacy of existing ATS evaluation metrics and further illustrates how human evaluation differs from existing evaluation systems.”
BackgroundResponsible AI Considerations in Text Summarization Research: A Review of Current Practices
Liu, Y. H., Cao, M., Blodgett, S. L., Cheung, J. C. K., Olteanu, A., & Trischler, A. (2023) · Findings of the Association for Computational Linguistics: EMNLP (Findings of EMNLP) · doi:10.18653/v1/2023.findings-emnlp.413 · Cited by 23
A Hybrid Intelligence Method for Argument Mining
van der Meer, M., Liscio, E., Jonker, C. M., Plaat, A., Vossen, P. T. J. M., & Murukannaiah, P. K. (2024) · Journal of Artificial Intelligence Research · doi:10.1613/jair.1.15135 · Cited by 13
“Specifically, BERT score recall correlates well with human consistency judgments, the factual alignment between selected argument and references (original opinions) (Fabbri et al., 2021).”
BackgroundHierarchical Indexing for Retrieval-Augmented Opinion Summarization
Hosking, T., Tang, H., & Lapata, M. (2024) · Transactions of the Association for Computational Linguistics · doi:10.1162/tacl_a_00703 · Cited by 13
Automatic Summarization of Russian Texts: Comparison of Extractive and Abstractive Methods
Головизнина, В. С., & Котельников, Е. В. (2022) · Computational Linguistics and Intellectual Technologies · doi:10.28995/2075-7182-2022-21-223-235 · Cited by 8
Unsupervised Extractive Summarization of Emotion Triggers
Sosea, T., Zhan, H., Li, J. J., & Caragea, C. (2023) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/2023.acl-long.531 · Cited by 11
Concise Answers to Complex Questions: Summarization of Long-form Answers
Potluri, A., Xu, F., & Choi, E. (2023) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/2023.acl-long.541 · Cited by 14
Revisiting text decomposition methods for NLI-based factuality scoring of summaries
Glover, J., Fancellu, F., Jagannathan, V., Gormley, M. R., & Schaaf, T. (2022) · Workshop on Generation, Evaluation and Metrics (GEM) · doi:10.18653/v1/2022.gem-1.7 · Cited by 17
Indian Legal NLP Benchmarks : A SurveyPreprint
Kalamkar, P., Venugopalan, J., & D, V. R. P. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2107.06056 · Cited by 13
“Recent research (Fabbri et al., 2021) proposes that there are 4 main aspects of summary evaluation: Coherence, Consistency , Flu-ency & Relevance.”
BackgroundCan GPT models Follow Human Summarization Guidelines? A Study for Targeted Communication GoalsPreprint
Zhou, Y., Ringeval, F., & Portet, F. (2023) · arXiv (Cornell University) · doi:10.48550/arxiv.2310.16810 · Cited by 10
“Automatic metrics show limited correlation with human judgment (Fabbri et al., 2021).”
Li, Y., Li, L., Hu, D., Li, Y., Zhou, Y., Litvak, M., & Vanetik, N. (2023) · Research Square · doi:10.21203/rs.3.rs-3695932/v1 · Cited by 10
RECON: Reasoning with Condensation for Efficient Retrieval-Augmented GenerationPreprint
Xu, Z., Wang, M., Wang, Y., Ye, W., Du, Y., Ma, Y., & Tian, Y. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2510.10448 · Cited by 9
“Together, these components enable efficient multi-turn retrieval and reasoning, while keeping context length under control.”
Balde, G., Roy, S., Mondal, M., & Ganguly, N. (2024) · Findings of the Association for Computational Linguistics: EMNLP 2024 · doi:10.18653/v1/2024.findings-emnlp.863 · Cited by 6
“We show the performance comparison results of A DAPT BPE in”
Hegselmann, S., Shen, S. Z., Gierse, F., Agrawal, M., Sontag, D., & Jiang, X. (2024) · arXiv · doi:10.48550/arxiv.2402.15422 · Cited by 6
“For our qualitative evaluation of patient summaries we used the dimension introduced in Fabbri et al. (2021).”
MethodKatzy, J., Huang, Y., Panchu, G.-R., Ziemlewski, M., Loizides, P., Vermeulen, S., van Deursen, A., & Izadi, M. (2025) · International Conference on Predictive Models in Software Engineering · doi:10.1145/3727582.3728683 · Cited by 9
EvalAssist: Insights on Task-Specific Evaluations and AI-Assisted Judgment Strategy Preferences
Ashktorab, Z., Desmond, M. J., Pan, Q., Johnson, J. M., Cooper, M. S., Daly, E., Nair, R., Pedapati, T., Do, H. J., & Geyer, W. (2025) · ACM Symposium on User Interface Software and Technology (UIST) · doi:10.1145/3746059.3747740 · Cited by 8
“For the summarization task, we presented the original reference document and corresponding generated outputs to users to be judged from [16].”
The Effect of Document Summarization on LLM-Based Relevance Judgments
Mohtadi, S., Roitero, K., Mizzaro, S., & Demartini, G. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-21300-6_5 · Cited by 7
ADSumm: annotated ground-truth summary datasets for disaster tweet summarization
Garg, P. K., Chakraborty, R., & Dandapat, S. K. (2024) · Social Network Analysis and Mining · doi:10.1007/s13278-024-01323-9 · Cited by 6
Cui, W., Li, Z., Lopez, D., Das, K., Malin, B., Kumar, S., & Zhang, J. (2024) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/2024.emnlp-industry.25 · Cited by 17
“Following the guidance of BERTScore (Zhang et al., 2020), we are using the PAWS development set and the first 5000 from the training set of QQP. • SummEval (Fabbri et al., 2021) is a standard dataset that assesses various summarization evaluation techniques.”
MethodCao, B., Guan, X., Bao, S., Wu, J., & Fan, J. (2025) · IEEE Transactions on Cognitive and Developmental Systems · doi:10.1109/tcds.2025.3563357 · Cited by 6
“Summaries are evaluated from three key perspectives: informativeness, quality [45], and coherence [46].”
Ferreira, A. A., da Rocha, L. C. D., da Cunha, W. L. M., Machado, A. C., Campos, J. M., Jallais, G. F., Viana, A. C. F., Tuler, E., Araujo, I. B., Macul, V. C., et al. (2025) · Scientific Reports · doi:10.1038/s41598-025-13560-9 · Cited by 9
“After reviewing the literature 3,25 and discussing the metrics with the Ana Health healthcare team, we focus our assessment on these perspectives.”
Predicting Intervention Approval in Clinical Trials through Multi-Document Summarization
Katsimpras, G., & Παλιούρας, Γ. (2022) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/2022.acl-long.137 · Cited by 6
Recent Advances in Chatbot Algorithms, Techniques, and Technologies: Designing Chatbots
Caldarini, G., & Jaf, S. (2023) · Advances in web technologies and engineering book series · doi:10.4018/978-1-6684-6234-8.ch011 · Cited by 6
Wang, Y., Wang, P., Zhang, Y., Yao, S., Xu, Z., & Zhang, Y. (2023) · Database · doi:10.1093/database/baad031 · Cited by 6
“After obtaining the initial filtered data, the long abstracts were compressed by Google’s pegasus-large model (27) for paragraph comprehension, information compression (28), and language generation (29) to produce an abstract overview of only a few sentences, the purpose of this step is to compress the long abstracts…”
MethodShen, X., Lam, W. P., Ma, S., & Wang, H. (2023) · Natural Language Engineering · doi:10.1017/s1351324923000177 · Cited by 7
Benchmarks for Indian Legal NLP: A Survey
Kalamkar, P., Venugopalan, J., & Raghavan, V. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-36190-6_3 · Cited by 6
Are Experts Needed? On Human Evaluation of Counselling Reflection Generation
Wu, Z., Balloccu, S., Reiter, E., Helaoui, R., Recupero, D. R., & Riboni, D. (2023) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/2023.acl-long.382 · Cited by 22
“Some human evaluation studies have compared expert and non-expert NLG evaluations, such as for summarisation (Gillick and Liu, 2010; Fabbri et al., 2021), machine translation (Freitag et al.”
BackgroundInducing Causal Structure for Abstractive Text Summarization
Chen, L., Zhang, R., Huang, W., Chen, W., Guo, J., & Cheng, X. (2023) · ACM International Conference on Information and Knowledge Management (CIKM) · doi:10.1145/3583780.3614934 · Cited by 6
“• Human Evaluation: We measure the Informativeness, Faithfulness, and Fluency referring to [13, 22, 27].”
MethodLength Does Matter: Summary Length can Bias Summarization Metrics
Guo, X., & Vosoughi, S. (2023) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/2023.emnlp-main.984 · Cited by 21
“Prior research, such as Yuan et al. (2021); Fabbri et al. (2021), highlighted variable average scores for some metrics depending on reference/generation lengths.”
BackgroundMethodInfluentialReference-Free Summarization Evaluation with Large Language Models
Akkasi, A., Fraser, K., & Komeili, M. (2023) · Workshop on Evaluation and Comparison of NLP Systems (Eval4NLP) · doi:10.18653/v1/2023.eval4nlp-1.16 · Cited by 12
“These summaries have been generated by a summarization model that was trained on the CNN/DailyMail dataset, as documented by (Fabbri et al., 2020).”
BackgroundMethodKotonya, N., Krishnasamy, S., Tetreault, J., & Jaimes, A. (2023) · Workshop on Evaluation and Comparison of NLP Systems (Eval4NLP) · doi:10.18653/v1/2023.eval4nlp-1.17 · Cited by 13
Wang, J., Zhang, S., Xu, Y., Zhang, Z., Yang, X., & Cheng, Y. (2025) · Expert Systems with Applications · doi:10.1016/j.eswa.2025.126592 · Cited by 6
Leveraging Requirements Elicitation through Software Requirement Patterns and LLMs
Franch, X., Gnesi, S., Paccosi, F., Quer, C., & Semini, L. (2025) · Lecture notes in computer science · doi:10.1007/978-3-031-88531-0_19 · Cited by 7
The current state of summarization
Retkowski, F. (2023) · AI critique · doi:10.1515/9783839467664-016 · Cited by 11
Text Alignment Is An Efficient Unified Model for Massive NLP Tasks
Zha, Y., Yang, Y., Li, R., & Hu, Z. (2023) · neural information processing systems · doi:10.52202/075280-3405 · Cited by 16
WIDAR - Weighted Input Document Augmented ROUGE
Jain, R., Mavi, V., Jangra, A., & Saha, S. (2022) · Lecture notes in computer science · doi:10.1007/978-3-030-99736-6_21 · Cited by 9
Deep Learning-Based Abstractive Summarization for Brazilian Portuguese Texts
Paiola, P. H., de Rosa, G. H., & Papa, J. P. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-21689-3_34 · Cited by 6
Ask, Retrieve, Summarize: A Modular Pipeline for Scientific Literature SummarizationPreprint
Achkar, P., Gollub, T., & Potthast, M. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2505.16349 · Cited by 7
“Experiments show that G-Eval achieves a stronger correlation with human judgments than traditional metrics such as ROUGE, as well as neural evaluators, on benchmarks such as SummEval [22] for summarization.”
Towards More Robust NLP System Evaluation: Handling Missing Scores in Benchmarks
Himmi, A., Irurozki, E., Noiry, N., Clemencon, S., & Colombo, P. (2023) · arXiv · doi:10.48550/arxiv.2305.10284 · Cited by 6
Large language models robustness against perturbation
Alahmari, S. S., Hall, L., Mouton, P. R., & Goldgof, D. (2025) · Scientific Reports · doi:10.1038/s41598-025-29770-0 · Cited by 5
“This work used two dataset, SummEval 23 and USR 24 , and showed that Google Gemini 1 falls short with the perturbations injected in the prompts.”
Zawistowski, K. (2024) · Annals of Computer Science and Information Systems · doi:10.15439/2024f7639 · Cited by 5
RetriEVAL: Evaluating Text Generation with Contextualized Lexical Match
Li, Z., Li, X., Tao, C., Feng, J., Shen, T., Xu, C., Wang, H., Zhao, D., & Ma, S. (2025) · ACM International Conference on Web Search and Data Mining (WSDM) · doi:10.1145/3701551.3703581 · Cited by 6
“In this paper, we conduct extensive experiments on two meta-evaluation benchmarks: SummEval [8] and Topical-Chat [30].”
CFAS: consensus-focused abstractive meeting summarization through multi-party discourse modeling
Jin, Y., Shi, Q., & Liu, Q. (2025) · Journal of King Saud University - Computer and Information Sciences · doi:10.1007/s44443-025-00210-3 · Cited by 5
“The next section formalizes our task and model, followed by details of each component (Lin et al. 2023; Koh et al. 2022; Fabbri et al. 2021). component and describe how they interact in an end-to-end framework.”
Ghosh, A., Jain, R., Jhangra, A., Saha, S., & Jatowt, A. (2025) · Wiley Interdisciplinary Reviews Data Mining and Knowledge Discovery · doi:10.1002/widm.70045 · Cited by 8
Motghare, M., Agarwal, M., & Agrawal, A. (2025) · Computers · doi:10.3390/computers14120508 · Cited by 5
Suharyadi, & Saputra, I. (2025) · Journal of Novel Engineering Science and Technology · doi:10.56741/jnest.v4i02.1042 · Cited by 5
“As emphasized by previous research [16][17][18], human judgment remains a critical benchmark in evaluating generative systems for complex domains.”
Mozaffari, N., Mehrangmarani, B., & Khorramrouz, A. (2026) · Human Behavior and Emerging Technologies · doi:10.1155/hbe2/8695447 · Cited by 8
HypoEval: Hypothesis-Guided Evaluation for Natural Language Generation
Li, M., Li, H., & Tan, C. (2026) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/2026.acl-long.1963 · Cited by 9
Anchoring Bias in LLM-as-a-Judge Systems: Prior Scores Compromise Evaluation Independence
Kapetanović, A. L., Altwlkany, K., Merćep, A., Đuričić, T., & Lacić, E. (2026) · ACM International Conference on Information and Knowledge Management (CIKM) · doi:10.1145/3799682.3840879 · Cited by 5
Deutsch, D., & Roth, D. (2022) · Findings of the Association for Computational Linguistics: ACL (Findings of ACL) · doi:10.18653/v1/2022.findings-acl.296 · Cited by 7
With a Little Push, NLI Models can Robustly and Efficiently Predict Faithfulness
Steen, J., Opitz, J., Frank, A., & Markert, K. (2023) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/2023.acl-short.79 · Cited by 17
Unsupervised Token-level Hallucination Detection from Summary Generation By-products
Marfurt, A., & Henderson, J. (2022) · Workshop on Generation, Evaluation and Metrics (GEM) · doi:10.18653/v1/2022.gem-1.21 · Cited by 13
“It consists of 250 summaries from the CNN/DailyMail test set, obtained from SummEval (Fabbri et al., 2021”
MethodBeyond Quantity: Research with Subsymbolic AI
Sudmann, A., Echterhölter, A., Ramsauer, M., Retkowski, F., Schröter, J., & Waibel, A. (2023) · AI critique · doi:10.1515/9783839467664 · Cited by 14
Large language models as partners in medical literature
Pérez-Guerrero, E. J., Mehrotra, I., Jain, S. S., & Perez, M. V. (2024) · Heart Rhythm · doi:10.1016/j.hrthm.2024.07.097 · Cited by 6
Beyond One-Size-Fits-All : Inversion Learning for Highly Effective NLG Evaluation Prompts
Hong, H., Xiao, C., Wang, Y., Liu, Y., Rong, W., & Lin, C. (2026) · Transactions of the Association for Computational Linguistics · doi:10.1162/tacl.a.617 · Cited by 13
Korenčić, D., & Grubišić, I. (2022) · International Workshop on Semantic Evaluation (SemEval) · doi:10.18653/v1/2022.semeval-1.5 · Cited by 6
Query-Utterance Attention With Joint Modeuing For Query-Focused Meeting Summarization
Liu, X., Duan, B., Xiao, B., & Xu, Y. (2023) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49357.2023.10096486 · Cited by 5
“Models were evaluated using the ROUGE metrics [9] in the SummEval toolkit [10] and each pair of results was subjected to t-test to confirm the effectiveness of our method.”
MethodNarrain, J. M., Taneja, V., Atrey, S. B., Sivaram, J., & Singh, D. (2023) · Winter Summit on Smart Computing and Networks (WiSSCoN) · doi:10.1109/wisscon56857.2023.10133863 · Cited by 5
“In their paper titled ’SummEval: Re-evaluating Summarization Evaluation’ [8], Fabbri, Kryściński et al.”
BackgroundMethodInfluentialFALTE: A Toolkit for Fine-grained Annotation for Long Text Evaluation
Goyal, T., Li, J. J., & Durrett, G. (2022) · Conference on Empirical Methods in Natural Language Processing: System Demonstrations (EMNLP) · doi:10.18653/v1/2022.emnlp-demos.35 · Cited by 10
“The study showed that the inter-annotator agreement of document-level labels for long texts is 0.19 compared to 0.48 reported for news summaries that are considerably shorter (Fabbri et al., 2021).”
BackgroundInfluentialMulti-Stage Pre-training Enhanced by ChatGPT for Multi-Scenario Multi-Domain Dialogue Summarization
Zhou, W., Li, G., Cheng, X., Liang, X., Zhu, J., Zhai, F., & Li, Z. (2023) · Findings of the Association for Computational Linguistics: EMNLP (Findings of EMNLP) · doi:10.18653/v1/2023.findings-emnlp.460 · Cited by 12
OpenAsp: A Benchmark for Multi-document Open Aspect-based Summarization
Amar, S., Schiff, L., Ernst, O., Shefer, A., Shapira, O., & Dagan, I. (2023) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/2023.emnlp-main.121 · Cited by 5
“Following Fabbri et al. (2021), a summary should be measured for 4 quality criteria: (1) relevance , the selection of important content from the source; (2) coherence , the quality of the collective structure of all sentences; (3) consistency , the factual alignment between the summary and the…”
BackgroundExploring Prompting Large Language Models as Explainable Metrics
Mahmoudi, G. (2023) · Workshop on Evaluation and Comparison of NLP Systems (Eval4NLP) · doi:10.18653/v1/2023.eval4nlp-1.18 · Cited by 5
Balde, G., Roy, S., Mondal, M., & Ganguly, N. (2024) · International Joint Conference on Artificial Intelligence · doi:10.24963/ijcai.2024/683 · Cited by 18
“…or Doctoral degree in Medicine and Biomedical Sciences and are older than 24 years) for annotating summary pairs of MEDVOC and IFT-CNN across the standard aspects [ Fabbri et al. , 2021b; Zhang et al. , 2023b ] of relevance , coherence (on a Likert scale of 1 to 5 ), and faithfulness (binary”
BackgroundMoFE: Mixture of Factual Experts for Controlling Hallucinations in Abstractive Summarization.Preprint
Choubey, P. K., Vig, J., Liu, W., & Rajani, N. F. (2021) · arXiv (Cornell University) · Cited by 6
Evaluating LLMs at Detecting Errors in LLM Responses
Kamoi, R., Das, S. S. S., Lou, R., Ahn, J. J., Zhao, Y., Lu, X., Zhang, N., Zhang, Y., Zhang, R. H., Vummanthala, S. R., et al. (2024) · arXiv · doi:10.48550/arxiv.2404.03602 · Cited by 5
“This criterion is often used as an objective “factuality” criterion for text summarization (Fabbri et al., 2021; Cao & Wang, 2021; Laban et al., 2023).”
BackgroundMethodUnderstanding Faithfulness and Reasoning of Large Language Models on Plain Biomedical Summaries
Fang, B., Dai, X., & Karimi, S. (2024) · Findings of the Association for Computational Linguistics: EMNLP 2024 · doi:10.18653/v1/2024.findings-emnlp.578 · Cited by 4
“Current faithfulness research on LLMs mostly focuses on the general domain, with a particular interest in news articles (Pagnoni et al., 2021; Fabbri et al., 2021; Tang et al., 2023; Cao and Wang, 2021).”
Alabdulwahab, A., Japic, C., Le, C., Dubey, D., Trivedi, D., Hope, J., Stone, P., Srivastava, S., Tashman, A. P., & Zhang, A. (2025) · Systems and Information Engineering Design Symposium · doi:10.1109/sieds65500.2025.11021089 · Cited by 6
“We used the SummEval [8] dataset that contains 100 CNN and Daily Mail articles.”
Benchmarking LLM-as-a-Judge Models for 5W1H Extraction Evaluation
Cassola-Bacallao, J. M., Morales-Donaire, J., Hernández-Montoya, P., & Keith-Norambuena, B. (2026) · Electronics · doi:10.3390/electronics15030659 · Cited by 5
Quantifying Factual Divergence in Generative Models: SHAP-LIME Based Hallucination Score for LLMs
Haq, I., Saqib, M., Zhang, Y., & Khan, I. A. (2026) · Multimedia Systems · doi:10.1007/s00530-025-02150-4 · Cited by 4
“Early hallucination detection efforts relied on heuristic and lexical similarity-based techniques, such as n-gram overlap, named-entity alignment, and source-reference matching [13].”
KEA Explain: A Neurosymbolic Framework for Explaining LLM Hallucinations
Haskins, R., & Adams, B. (2026) · Neurosymbolic Artificial Intelligence · doi:10.1177/29498732261443099 · Cited by 5
Metadata
- Authors
- Fabbri, A. R., Kryściński, W., McCann, B., Xiong, C., Socher, R., & Radev, D.
- Affiliations
- Yale University (US)
- Salesforce (United States) (US)
- Venue
- Transactions of the Association for Computational Linguistics
- Publisher
- MIT Press
- Published
- 2021
- Volume
- 9
- Pages
- 391-409
- ISSN
- 2307-387X (ISSN-L)
- OpenAlex
- W3045321166
- Cited by
- 587
- References
- 102
- Citations by year
- 2026: 109 · 2025: 93 · 2024: 74 · 2023: 166 · 2022: 84 · 2021: 57
- FWCI
- 33.37
- Type
- journal article
- Language
- en
- Open access
- diamond
- doi.org
- Locations
- Transactions of the Association for Computational Linguistics (publishedVersion, cc-by)
- arXiv (Cornell University) (submittedVersion)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
- arXiv (Cornell University)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Topic Modeling
- Physical Sciences › Computer Science › Artificial Intelligence › Advanced Text Analysis Techniques
- Social Sciences › Linguistics and Language › Text Readability and Simplification
- Keywords
- text summarization, summarization evaluation, automatic evaluation metrics, crowdsourced annotation, evaluation toolkit
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.