Spurious Correlations in Reference-Free Evaluation of Text Generation
Open accessAbstract
Model-based, reference-free evaluation metrics have been proposed as a fast and cost-effective approach to evaluate Natural Language Generation (NLG) systems.Despite promising recent results, we find evidence that reference-free evaluation metrics of summarization and dialog generation may be relying on spurious correlations with measures such as word overlap, perplexity, and length.We further observe that for text summarization, these metrics have high error rates when ranking current state-ofthe-art abstractive summarization systems.We demonstrate that these errors can be mitigated by explicitly designing evaluation metrics to avoid spurious features in reference-free evaluation.
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
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?
Perplexity is used in some long dialogue benchmarks as a likelihood-based metric for generation quality over sessions, but it remains an indirect indicator for memory performance, because it does not verify whether the model's content is grounded in the correct historical evidence.
“Perplexity is also used in some long dialogue benchmarks as a likelihood-based metric for generation quality over sessions (Xu et al., 2022a), but it remains an indirect indicator for memory performance, because it does not verify whether the model's content is grounded in the correct historical evidence (Durmus et al., 2022).”
From A Survey of Agent Memory in the Second Half: Towards Self-Evolving and Long-Horizon Agents (2026) · cites this paper as Durmus et al., 2022
Split into 2 atoms — 0 confirmed, 2 partially supported, 0 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.
Experimental — accuracy not yet measured
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
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 · see citing works
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 · see citing works
Transformers: 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 · see citing works
Domain-Adversarial Training of Neural Networks
Ganin, Y., Ustinova, E., Ajakan, H., Germain, P., Larochelle, H., Laviolette, F., Marchand, M., & Lempitsky, V. (2017) · Advances in Computer Vision and Pattern Recognition · doi:10.1007/978-3-319-58347-1_10 · Cited by 8,318 · see citing works
Get To The Point: Summarization with Pointer-Generator Networks
See, A., Liu, P. J., & Manning, C. D. (2017) · Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · doi:10.18653/v1/p17-1099 · Cited by 3,775 · see citing works
ELECTRA: Pre-training Text Encoders as Discriminators Rather Than Generators
Clark, K., Luong, M.-T., Le, Q. V., & Manning, C. D. (2020) · arXiv · doi:10.48550/arxiv.2003.10555 · Cited by 1,844 · see citing works
Liu, C., Lowe, R., Serban, I. V., Noseworthy, M., Charlin, L., & Pineau, J. (2016) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/d16-1230 · Cited by 1,367 · see citing works
PEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarization
Zhang, J., Zhao, Y., Saleh, M., & Liu, P. J. (2019) · arXiv · doi:10.48550/arxiv.1912.08777 · Cited by 1,206 · see citing works
Personalizing Dialogue Agents: I have a dog, do you have pets too?
Zhang, S., Dinan, E., Urbanek, J., Szlam, A., Kiela, D., & Weston, J. (2018) · Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · doi:10.18653/v1/p18-1205 · Cited by 1,177 · see citing works
DailyDialog: A Manually Labelled Multi-turn Dialogue DatasetPreprint
Li, Y., Su, H., Shen, X., Li, W., Cao, Z., & Niu, S. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1710.03957 · Cited by 757 · see citing works
Bottom-Up Abstractive Summarization
Gehrmann, S., Deng, Y., & Rush, A. M. (2018) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/d18-1443 · Cited by 679 · see citing works
Evaluating the Factual Consistency of Abstractive Text Summarization
Kryscinski, W., McCann, B., Xiong, C., & Socher, R. (2020) · Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/2020.emnlp-main.750 · Cited by 669 · see citing works
Wizard of Wikipedia: Knowledge-Powered Conversational agentsPreprint
Dinan, E., Roller, S., Shuster, K., Fan, A., Auli, M., & Weston, J. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1811.01241 · Cited by 610 · see citing works
Fast Abstractive Summarization with Reinforce-Selected Sentence Rewriting
Chen, Y.-C., & Bansal, M. (2018) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/p18-1063 · Cited by 601 · see citing works
SummEval: Re-evaluating Summarization Evaluation
Fabbri, A. R., Kryściński, W., McCann, B., Xiong, C., Socher, R., & Radev, D. (2021) · Transactions of the Association for Computational Linguistics · doi:10.1162/tacl_a_00373 · Cited by 587 · see citing works
Newsroom: A Dataset of 1.3 Million Summaries with Diverse Extractive Strategies
Grusky, M., Naaman, M., & Artzi, Y. (2018) · Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT) · doi:10.18653/v1/n18-1065 · Cited by 491 · see citing works
Fine-Tuning Language Models from Human Preferences
Ziegler, D. M., Stiennon, N., Wu, J., Brown, T. B., Radford, A., Amodei, D., Christiano, P., & Irving, G. (2019) · arXiv · doi:10.48550/arxiv.1909.08593 · Cited by 413 · see citing works
Faithful to the Original: Fact Aware Neural Abstractive Summarization
Cao, Z., Wei, F., Li, W., & Li, S. (2018) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v32i1.11912 · Cited by 359 · see citing works
(2020) · doi:10.18653/v1/2020.acl-main.450 · Cited by 356 · see citing works
Neural Document Summarization by Jointly Learning to Score and Select Sentences
Zhou, Q. S., Yang, N., Wei, F., Huang, S., Zhou, M., & Zhao, T. (2018) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/p18-1061 · Cited by 355 · see citing works
Durmus, E., He, H., & Diab, M. (2020) · Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics · doi:10.18653/v1/2020.acl-main.454 · Cited by 306 · see citing works
Self-citationFalke, T., Ribeiro, L. F. R., Utama, P. A., Dagan, I., & Gurevych, I. (2019) · Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics · doi:10.18653/v1/p19-1213 · Cited by 265 · see citing works
A Unified Model for Extractive and Abstractive Summarization using Inconsistency Loss
Hsu, W.-T., Lin, C.-K., Lee, M.-Y., Min, K., Tang, J. P., & Sun, M. (2018) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/p18-1013 · Cited by 253 · see citing works
Topical-Chat: Towards Knowledge-Grounded Open-Domain Conversations
Gopalakrishnan, K., Hedayatnia, B., Chen, Q., Gottardi, A., Kwatra, S., Venkatesh, A., Gabriel, R., & Hakkani-Tür, D. (2019) · Interspeech 2019 · doi:10.21437/interspeech.2019-3079 · Cited by 207 · see citing works
Evaluation of Text Generation: A Survey
Celikyilmaz, A., Clark, E., & Gao, J. (2020) · arXiv · doi:10.48550/arxiv.2006.14799 · Cited by 207 · see citing works
Tangled up in BLEU: Reevaluating the Evaluation of Automatic Machine Translation Evaluation Metrics
Mathur, N., Baldwin, T., & Cohn, T. (2020) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/2020.acl-main.448 · Cited by 206 · see citing works
Understanding Factuality in Abstractive Summarization with FRANK: A Benchmark for Factuality Metrics
Pagnoni, A., Balachandran, V., & Tsvetkov, Y. (2021) · Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies · doi:10.18653/v1/2021.naacl-main.383 · Cited by 197 · see citing works
BanditSum: Extractive Summarization as a Contextual Bandit
Dong, Y., Shen, Y., Crawford, E., van Hoof, H., & Cheung, J. C. K. (2018) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/d18-1409 · Cited by 171 · see citing works
Multi-Reward Reinforced Summarization with Saliency and Entailment
Pasunuru, R., & Bansal, M. (2018) · Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT) · doi:10.18653/v1/n18-2102 · Cited by 166 · see citing works
Improving Abstraction in Text Summarization
Kryściński, W., Paulus, R., Xiong, C., & Socher, R. (2018) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/d18-1207 · Cited by 163 · see citing works
The GEM Benchmark: Natural Language Generation, its Evaluation and Metrics
Gehrmann, S., Adewumi, T., Aggarwal, K., Ammanamanchi, P. S., Aremu, A., Bosselut, A., Chandu, K. R., Clinciu, M.-A., Das, D., Dhole, K., et al. (2021) · Proceedings of the 1st Workshop on Natural Language Generation, Evaluation, and Metrics (GEM 2021) · doi:10.18653/v1/2021.gem-1.10 · Cited by 159 · see citing works
Self-citationRe-evaluating Evaluation in Text Summarization
Bhandari, M., Gour, P. N., Ashfaq, A., Liu, P., & Neubig, G. (2020) · Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/2020.emnlp-main.751 · Cited by 148 · see citing works
Soft Layer-Specific Multi-Task Summarization with Entailment and Question Generation
Guo, H., Pasunuru, R., & Bansal, M. (2018) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/p18-1064 · Cited by 141 · see citing works
Learning to Extract Coherent Summary via Deep Reinforcement Learning
Wu, Y., & Hu, B. (2018) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v32i1.11987 · Cited by 128 · see citing works
To Ship or Not to Ship: An Extensive Evaluation of Automatic Metrics for Machine TranslationPreprint
Kocmi, T., Federmann, C., Grundkiewicz, R., Junczys-Dowmunt, M., Matsushita, H., & Menezes, A. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2107.10821 · Cited by 114 · see citing works
Annotating and Modeling Fine-grained Factuality in Summarization
Goyal, T., & Durrett, G. (2021) · Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies · doi:10.18653/v1/2021.naacl-main.114 · Cited by 100 · see citing works
Evaluating Factuality in Generation with Dependency-level Entailment
Goyal, T., & Durrett, G. (2020) · Findings of the Association for Computational Linguistics: EMNLP 2020 · doi:10.18653/v1/2020.findings-emnlp.322 · Cited by 93 · see citing works
Dialogue Response Ranking Training with Large-Scale Human Feedback Data
Gao, X., Zhang, Y., Galley, M., Brockett, C., & Dolan, B. (2020) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/2020.emnlp-main.28 · Cited by 84 · see citing works
GO FIGURE: A Meta Evaluation of Factuality in Summarization
Gabriel, S., Celikyilmaz, A., Jha, R., Choi, Y., & Gao, J. (2021) · Findings of the Association for Computational Linguistics: ACL-IJCNLP 2021 · doi:10.18653/v1/2021.findings-acl.42 · Cited by 75 · see citing works
A Comprehensive Assessment of Dialog Evaluation Metrics
Yeh, Y.-T., Eskénazi, M., & Mehri, S. (2021) · doi:10.18653/v1/2021.eancs-1.3 · Cited by 71 · see citing works
Evaluating Attribution in Dialogue Systems: The BEGIN Benchmark
Dziri, N., Rashkin, H., Linzen, T., & Reitter, D. (2022) · Transactions of the Association for Computational Linguistics · doi:10.1162/tacl_a_00506 · Cited by 67 · see citing works
Learning an Unreferenced Metric for Online Dialogue Evaluation
Sinha, K., Parthasarathi, P., Wang, J., Lowe, R., Hamilton, W. L., & Pineau, J. (2020) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/2020.acl-main.220 · Cited by 64 · see citing works
Evaluating 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 · see citing works
Ladhak, F., Durmus, E., He, H., Cardie, C., & McKeown, K. (2022) · Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · doi:10.18653/v1/2022.acl-long.100 · Cited by 46 · see citing works
Self-citationDeconstruct to Reconstruct a Configurable Evaluation Metric for Open-Domain Dialogue Systems
Phy, V., Zhao, Y., & Aizawa, A. (2020) · International Conference on Computational Linguistics · doi:10.18653/v1/2020.coling-main.368 · Cited by 39 · see citing works
Closed-Book Training to Improve Summarization Encoder Memory
Jiang, Y., & Bansal, M. (2018) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/d18-1440 · Cited by 34 · see citing works
Curious Case of Language Generation Evaluation Metrics: A Cautionary Tale
Çağlayan, O., Madhyastha, P., & Specia, L. (2020) · International Conference on Computational Linguistics · doi:10.18653/v1/2020.coling-main.210 · Cited by 28 · see citing works
SummEval: Re-evaluating Summarization Evaluation
Fabbri, A. R., Kryściński, W., McCann, B., et al. (2020) · arXiv · doi:10.48550/arxiv.2007.12626
Ladhak, F., Durmus, E., He, H., Cardie, C., & McKeown, K. (2021) · arXiv · doi:10.48550/arxiv.2108.13684
Self-citation
Flagged first, then by citations
Cited by
Sorted by citations
Benchmarking Large Language Models for News Summarization
Zhang, T., Ladhak, F., Durmus, E., Liang, P., McKeown, K., & Hashimoto, T. B. (2024) · Transactions of the Association for Computational Linguistics · doi:10.1162/tacl_a_00632 · Cited by 415 · see citing works
Self-citationRepairing the Cracked Foundation: A Survey of Obstacles in Evaluation Practices for Generated Text
Gehrmann, S., Clark, E., & Sellam, T. (2023) · Journal of Artificial Intelligence Research · doi:10.1613/jair.1.13715 · Cited by 123 · see citing works
Evaluating Attribution in Dialogue Systems: The BEGIN Benchmark
Dziri, N., Rashkin, H., Linzen, T., & Reitter, D. (2022) · Transactions of the Association for Computational Linguistics · doi:10.1162/tacl_a_00506 · Cited by 67 · see citing works
On the Limitations of Reference-Free Evaluations of Generated Text
Deutsch, D., Dror, R., & Roth, D. (2022) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/2022.emnlp-main.753 · Cited by 32 · see citing works
Federated Topic-Preference Learning for Knowledge-Grounded Chat with Differential Privacy
Kuo, M.-J., Zheng, D., & Hires, J. (2025) · Journal of Technology Informatics and Engineering · doi:10.51903/jtie.v4i2.502 · Cited by 27 · see citing works
Opportunities and Risks of LLMs for Scalable Deliberation with Polis
Small, C. T., Vendrov, I., Durmus, E., Homaei, H., Barry, E., Cornebise, J., Suzman, T., Ganguli, D., & Megill, C. (2023) · arXiv · doi:10.48550/arxiv.2306.11932 · Cited by 20 · see citing works
Self-citationLee, S. Y.-T., Bahukhandi, A., Liu, D., & Ma, K. (2024) · IEEE Transactions on Visualization and Computer Graphics · doi:10.1109/tvcg.2024.3456398 · Cited by 19 · see citing works
Length-Controlled AlpacaEval: A Simple Way to Debias Automatic Evaluators
Dubois, Y., Galambosi, B., Liang, P., & Hashimoto, T. B. (2024) · arXiv · doi:10.48550/arxiv.2404.04475 · Cited by 17 · see citing works
Self-citationA Meta-Evaluation of Faithfulness Metrics for Long-Form Hospital-Course Summarization
Adams, G., Zucker, J. E., & Elhadad, N. (2023) · PubMed · doi:10.48550/arxiv.2303.03948 · Cited by 15 · see citing works
Adams, G., Nguyen, B. H., Smith, J. A., Xia, Y., Xie, S., Ostropolets, A., Deb, B., Chen, Y.-J., Naumann, T., & Elhadad, N. (2023) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/2023.acl-long.587 · Cited by 9 · see citing works
Benchmarking and Assessing Transformations Between Text and Causal Maps via Large Language Models
Giabbanelli, P. J., Gandee, T. J., Agrawal, A., & Hosseinichimeh, N. (2025) · Applied Ontology · doi:10.1177/15705838241304102 · Cited by 8 · see citing works
SESCORE2: Learning Text Generation Evaluation via Synthesizing Realistic Mistakes
Xu, W., Qian, X., Wang, M., Li, L., & Wang, W. Y. (2023) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/2023.acl-long.283 · Cited by 7 · see citing works
On the shortcut learning in multilingual neural machine translation
Wang, W., Jiao, W., Huang, J.-T., Tu, Z., & Lyu, M. R.-T. (2024) · Neurocomputing · doi:10.1016/j.neucom.2024.128833 · Cited by 5 · see citing works
What Is Your Favorite Gender, MLM? Gender Bias Evaluation in Multilingual Masked Language Models
Yu, J., Kim, S. U., Choi, J., & Choi, J. D. (2024) · Information · doi:10.3390/info15090549 · Cited by 5 · see citing works
A Study on Automatic Open-Domain Dialogue Evaluation Metrics
Kim, M., & Kim, J. (2024) · International Conference on Consumer Electronics · doi:10.1109/icce59016.2024.10444444 · Cited by 2 · see citing works
On the Shortcut Learning in Multilingual Neural Machine Translation
Wang, W., Jiao, W., Huang, J.-T., Tu, Z., & Lyu, M. R. (2023) · doi:10.2139/ssrn.4474513 · Cited by 1 · see citing works
A Survey of Agent Memory in the Second Half: Towards Self-Evolving and Long-Horizon Agents
Huang, W.-C., Zhang, W., Liang, Y., Bei, Y., Chen, Y., Feng, T., Pan, X., Tan, Z., Wang, Y., Wei, T., et al. (2026) · arXiv · doi:10.48550/arxiv.2602.06052
On the Shortcut Learning in Multilingual Neural Machine TranslationPreprint
Wang, W., Jiao, W., Huang, J.-T., Tu, Z., & Lyu, M. R.-T. (2024) · SSRN Electronic Journal · doi:10.2139/ssrn.4862335
MAGNET: Counterfactual samples synthesizing for mitigating hallucination in large language models
Kim, B. S., Kim, B., & Jang, B. (2026) · PLoS ONE · doi:10.1371/journal.pone.0340812
What Stays and What Goes: Auditing the Impact of LLM Summarization on News Partisanship
Savgira, P., Kreiss, E., & Hosseinmardi, H. (2026) · CHI Conference Extended Abstracts on Human Factors in Computing Systems (CHI EA) · doi:10.1145/3772363.3799057
- Authors
- Durmus, E., Ladhak, F., & Hashimoto, T.
- Affiliations
- Stanford University (US)
- Columbia University (US)
- Venue
- Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
- Published
- 2022
- Pages
- 1443-1454
- ISSN
- 0736-587X (ISSN-L)
- OpenAlex
- W4224280227
- Cited by
- 16
- References
- 46
- Citations by year
- 2026: 2 · 2025: 1 · 2024: 6 · 2023: 5 · 2022: 2
- FWCI
- 1.94
- Type
- conference paper
- Language
- en
- Open access
- hybrid
- doi.org
- Licence
- cc-by
- Locations
- Proceedings of the conference - Association for Computational Linguistics. Meeting (publishedVersion, cc-by)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Natural Language Processing Techniques
- Physical Sciences › Computer Science › Artificial Intelligence › Topic Modeling
- Physical Sciences › Computer Science › Artificial Intelligence › Speech and dialogue systems
- Keywords
- reference-free evaluation, spurious correlations, natural language generation evaluation, text summarization, abstractive summarization, dialogue generation, lexical overlap, perplexity, model-based evaluation
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
