Investigating Educational and Noneducational Answer Selection for Educational Question Generation
Open accessAbstract
Educational automatic question generation (AQG) is often unable to realize its full potential in educational applications due to insufficient training data. For this reason, current research relies on noneducational question answering datasets for system training and evaluation. However, noneducational training data may comprise different language patterns than educational data. Consequently, the research question of whether models trained on noneducational datasets transfer well to the educational AQG task arises. In this work, we investigate the AQG subtask of answer selection, which aims to extract meaningful answers for the questions to be generated. We train and evaluate six modern and well-established BERT-based machine learning model architectures on two widely used noneducational datasets. Furthermore, we introduce a novel, midsized educational dataset for answer selection called TQA-A, which is used to investigate the transfer capabilities of the noneducational models to the educational domain. In terms of phrase-level evaluation metrics, noneducational models perform similar to models trained directly on the novel educational TQA-A dataset, although trained with considerably more training data. Moreover, models trained directly on TQA-A select fewer named entity-based and more verb-based answers than noneducational models. This provides evidence for differences in noneducational and educational answer selection tasks.
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?
Some applications used datasets such as SQuAD and MS-MARCO, while some works curated their own datasets for distinctive applications.
“For a few of these applications, some datasets were used like SQuAD, MS-MARCO and the like. Some works curated their own dataset owing to the distinctive application involved [122, 124].”
From Automatic question generation: a review of methodologies, datasets, evaluation metrics, and applications (2023) · cites this paper as [124]
[124] found a significant difference in how answers were selected in educational versus non-educational question generation.
“This study found that there was a significant difference in which the answers were selected in educational and non-educational question generation.”
From Automatic question generation: a review of methodologies, datasets, evaluation metrics, and applications (2023) · cites this paper as [124]
Standalone question generation has been applied to education, news, and social media.
“If we consider the different domains, standalone question generation has been applied to education [123], news [124] and social media [126].”
From Automatic question generation: a review of methodologies, datasets, evaluation metrics, and applications (2023) · cites this paper as [124]
In [124], transfer learning trains on non-educational datasets SQuAD and NQA and evaluates on the author-curated TQA-A dataset of educational text with answer tags.
“For instance, in [124], transfer learning is performed by training on non-educational datasets like SQuAD and NQA (Natural Questions) and the evaluation is performed on an author curated dataset called TQA-A with questions based on educational text and tagged with answers.”
From Automatic question generation: a review of methodologies, datasets, evaluation metrics, and applications (2023) · cites this paper as [124]
Most works used factoid-based questions, but Yes/No, MCQ, and reasoning-type questions were also generated for some applications.
“Also, most works used factoid-based questions but Yes/No [133], MCQ type [124] and reasoning type questions [122] were also generated for some applications.”
From Automatic question generation: a review of methodologies, datasets, evaluation metrics, and applications (2023) · cites this paper as [124]
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
BERT: 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 48,293 · cited 2× · see citing works
“Hence, we deemed six epochs as sufficient for the task, in line with findings by the BERT authors, suggesting that four epochs are sufficient for various fine-tuning tasks [13].”
Method(2008) · Journal of Machine Learning Research · Cited by 33,051 · see citing works
Efficient Estimation of Word Representations in Vector Space
Mikolov, T., Chen, K., Corrado, G., & Dean, J. (2013) · arXiv · doi:10.48550/arxiv.1301.3781 · Cited by 26,568 · cited 2× · see citing works
“Each sentence embedding vector is a Continuous Bag of Words [11] averaging over all word embeddings of the sentence 3 .”
BackgroundRoBERTa: A Robustly Optimized BERT Pretraining Approach
Liu, Y., Ott, M., Goyal, N., Du, J., Joshi, M., Chen, D., Levy, O., Lewis, M., Zettlemoyer, L., & Stoyanov, V. (2019) · arXiv · doi:10.48550/arxiv.1907.11692 · Cited by 16,515 · cited 3× · see citing works
“Second, ROBERTa [14] is included in the comparison.”
BackgroundMethodSQuAD: 100,000+ Questions for Machine Comprehension of Text
Rajpurkar, P., Zhang, J., Lopyrev, K., & Liang, P. (2016) · Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing · doi:10.18653/v1/d16-1264 · Cited by 6,208 · cited 4× · see citing works
“Among these commonly used question answering data-sets are SQuAD [5] and Natural Questions (NQA) [7].”
BackgroundMethodInfluentialDistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter
Sanh, V., Debut, L., Chaumond, J., & Wolf, T. (2019) · arXiv · doi:10.48550/arxiv.1910.01108 · Cited by 4,653 · cited 2× · see citing works
“Additionally, we investigate DistilBERT [16], which has a magnitude fewer parameters than other BERT architectures but has nevertheless achieved similar results on a variety of tasks [16].”
MethodResultALBERT: A Lite BERT for Self-supervised Learning of Language Representations
Lan, Z., Chen, M., Goodman, S., Gimpel, K., Sharma, P., & Soricut, R. (2019) · arXiv · doi:10.48550/arxiv.1909.11942 · Cited by 4,308 · cited 8× · see citing works
“Six different architectures are investigated: BERT [13], RoBERTa [14], DEBERTa [15], DistilBERT [16], ALBERT [17], and SpanBERT [18].”
BackgroundMethodInfluentialGLUE: A Multi-Task Benchmark and Analysis Platform for Natural Language Understanding
Wang, A., Singh, A., Michael, J., Hill, F., Levy, O., & Bowman, S. (2018) · Proceedings of the 2018 EMNLP Workshop BlackboxNLP: Analyzing and Interpreting Neural Networks for NLP · doi:10.18653/v1/w18-5446 · Cited by 4,222 · cited 1× · see citing works
“It contributes a significantly better model and, for instance, outperforms the BERT model on the important GLUE [19] benchmark [14].”
MethodHuggingFace's 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. (2019) · arXiv · doi:10.48550/arxiv.1910.03771 · Cited by 3,347 · see citing works
Natural Questions: A benchmark for question answering research
Kwiatkowski, T., Palomaki, J., Redfield, O., Collins, M., Parikh, A., Alberti, C., Epstein, D., Polosukhin, I., Devlin, J., Lee, K., et al. (2019) · Transactions of the Association for Computational Linguistics · doi:10.1162/tacl_a_00276 · Cited by 2,069 · see citing works
SpanBERT: Improving Pre-training by Representing and Predicting Spans
Joshi, M., Chen, D., Liu, Y., Weld, D. S., Zettlemoyer, L., & Levy, O. (2020) · Transactions of the Association for Computational Linguistics · doi:10.1162/tacl_a_00300 · Cited by 1,827 · cited 2× · see citing works
“Its parameter size is comparable to BERT, but it is explicitly designed toward text span prediction [18].”
MethodDEBERTA: DECODING-ENHANCED BERT WITH DISENTANGLED ATTENTION
He, P., Liu, X., Gao, J., & Chen, W. (2021) · International Conference on Learning Representations · Cited by 1,116 · see citing works
RACE: Large-scale ReAding Comprehension Dataset From Examinations
Lai, G., Xie, Q., Liu, H., Yang, Y., & Hovy, E. (2017) · Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing · doi:10.18653/v1/d17-1082 · Cited by 1,044 · cited 2× · see citing works
“However, all three datasets [10], [21], [22] only contain a document and corresponding questions and do not annotate the answer span inside the document explicitly.”
BackgroundLearning to Ask: Neural Question Generation for Reading Comprehension
Du, X., Shao, J., & Cardie, C. (2017) · Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · doi:10.18653/v1/p17-1123 · Cited by 696 · cited 3× · see citing works
“The SQuAD dataset is probably the most widespread dataset for AQG, and it has been shown that systems trained on the dataset generate fluent questions useful for augmenting question answering training data (e.g. [3], [8]).”
BackgroundMethodSQuAD: 100,000+ Questions for Machine Comprehension of Text
Rajpurkar, P., Zhang, J., Lopyrev, K., & Liang, P. (2016) · arXiv · doi:10.48550/arxiv.1606.05250 · Cited by 678 · cited 4× · see citing works
“Among these commonly used question answering data-sets are SQuAD [5] and Natural Questions (NQA) [7].”
BackgroundMethodInfluentialGLUE: A Multi-Task Benchmark and Analysis Platform for Natural Language Understanding
Wang, A., Singh, A., Michael, J., Hill, F., Levy, O., & Bowman, S. R. (2018) · arXiv · doi:10.48550/arxiv.1804.07461 · Cited by 562 · cited 1× · see citing works
“It contributes a significantly better model and, for instance, outperforms the BERT model on the important GLUE [19] benchmark [14].”
MethodProphetNet: Predicting Future N-gram for Sequence-to-SequencePre-training
Qi, W., Yan, Y., Gong, Y., Liu, D., Duan, N., Chen, J., Zhang, R., & Zhou, M. (2020) · Findings of the Association for Computational Linguistics: EMNLP (Findings of EMNLP) · doi:10.18653/v1/2020.findings-emnlp.217 · Cited by 401 · see citing works
Kembhavi, A., Seo, M., Schwenk, D., Choi, J., Farhadi, A., & Hajishirzi, H. (2017) · 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2017.571 · Cited by 244 · cited 3× · see citing works
“However, all three datasets [10], [21], [22] only contain a document and corresponding questions and do not annotate the answer span inside the document explicitly.”
BackgroundHarvesting Paragraph-level Question-Answer Pairs from Wikipedia
Du, X., & Cardie, C. (2018) · Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · doi:10.18653/v1/p18-1177 · Cited by 168 · cited 2× · see citing works
“The supervised methods usually rely on neural networks [6], [27], [28] and model the task as sequence labelling [6], [27] or classification task [29], [30].”
BackgroundMethodRACE: Large-scale ReAding Comprehension Dataset From ExaminationsPreprint
Lai, G., Xie, Q., Liu, H., et al. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1704.04683 · Cited by 108 · cited 2× · see citing works
“However, all three datasets [10], [21], [22] only contain a document and corresponding questions and do not annotate the answer span inside the document explicitly.”
BackgroundInvestigating Transformers for Automatic Short Answer Grading
Camus, L., & Filighera, A. (2020) · Lecture notes in computer science · doi:10.1007/978-3-030-52240-7_8 · Cited by 87 · cited 1× · see citing works
“Fields like multimodal learning analytics (MM-LA) or automatic short answer grading (ASAG) show promising use cases for artificial intelligence-powered applications [1], [2].”
BackgroundSelf-citationLearningQ: A Large-Scale Dataset for Educational Question Generation
Chen, G., Yang, J., Hauff, C., & Houben, G.-J. (2018) · Proceedings of the International AAAI Conference on Web and Social Media · doi:10.1609/icwsm.v12i1.14987 · Cited by 83 · cited 3× · see citing works
“However, all three datasets [10], [21], [22] only contain a document and corresponding questions and do not annotate the answer span inside the document explicitly.”
BackgroundMethodWang, Z., Lan, A., Nie, W., Waters, A. E., Grimaldi, P. J., & Baraniuk, R. G. (2018) · ACM Conference on Learning @ Scale (L@S) · doi:10.1145/3231644.3231654 · Cited by 71 · cited 1× · see citing works
“Furthermore, Steuer et al. [25], and Wang et al. [26] evaluated AQG systems trained on SQuAD on educational corpora with promising results.”
BackgroundRevUP: Automatic Gap-Fill Question Generation from Educational Texts
Kumar, G., Banchs, R. E., & D’Haro, L. F. (2015) · Workshop on Innovative Use of NLP for Building Educational Applications · doi:10.3115/v1/w15-0618 · Cited by 51 · cited 1× · see citing works
“The supervised methods usually rely on neural networks [6], [27], [28] and model the task as sequence labelling [6], [27] or classification task [29], [30].”
MethodMind the Gap: Learning to Choose Gaps for Question Generation
Becker, L. A., Basu, S., & Vanderwende, L. (2012) · Cited by 49 · see citing works
Automating Reading Comprehension by Generating Question and Answer Pairs
Kumar, V., Boorla, K., Meena, Y. K., Ramakrishnan, G., & Li, Y.-F. (2018) · Lecture notes in computer science · doi:10.1007/978-3-319-93040-4_27 · Cited by 49 · see citing works
Key Phrase Extraction for Generating Educational Question-Answer Pairs
Willis, A., Davis, G. M., Ruan, S., Manoharan, L., Landay, J. A., & Brunskill, E. (2019) · ACM Conference on Learning @ Scale (L@S) · doi:10.1145/3330430.3333636 · Cited by 41 · cited 4× · see citing works
“Willis et al. [27] address the problem using generative models.”
BackgroundMethodInfluentialInfusing NLU into Automatic Question Generation
Mazidi, K., & Tarau, P. (2016) · Proceedings of the 9th International Natural Language Generation conference · doi:10.18653/v1/w16-6609 · Cited by 36 · cited 1× · see citing works
“Furthermore, unsupervised approaches have been investigated [33] V. DISCUSSION We have investigated the performance of six answer selection models on three different datasets and their transferability to the educational domain.”
BackgroundQu, F., Jia, X., & Wu, Y. (2021) · Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing · doi:10.18653/v1/2021.emnlp-main.202 · Cited by 30 · cited 1× · see citing works
“Nevertheless, these datasets have been used to build automatic question generators [22], [23] although the evaluation did not measure the validity of answer selection approaches directly.”
MethodNeural Question Generation with Answer Pivot
Wang, B., Wang, X., Tao, T., Zhang, Q., & Xu, J. (2020) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v34i05.6449 · Cited by 27 · cited 1× · see citing works
“The SQuAD dataset is probably the most widespread dataset for AQG, and it has been shown that systems trained on the dataset generate fluent questions useful for augmenting question answering training data (e.g. [3], [8]).”
MethodReal-Time Multimodal Feedback with the CPR Tutor
Mitri, D. D., Schneider, J., Trebing, K., Sopka, S., Specht, M. M., & Drachsler, H. (2020) · Lecture notes in computer science · doi:10.1007/978-3-030-52237-7_12 · Cited by 27 · see citing works
Automatically Generating Cause-and-Effect Questions from Passages
Stasaski, K., Rathod, M., Tu, T., Xiao, Y., & Hearst, M. A. (2021) · Workshop on Innovative Use of NLP for Building Educational Applications · Cited by 19 · see citing works
Chen, G., Yang, J., & Gašević, D. (2019) · Lecture notes in computer science · doi:10.1007/978-3-030-23204-7_6 · Cited by 18 · cited 1× · see citing works
“Besides direct educational data collection, some works have investigated the transfer of machine learning approaches on noneducational data to the educational tasks [24]–[27].”
BackgroundRemember the Facts? Investigating Answer-Aware Neural Question Generation for Text Comprehension
Steuer, T., Filighera, A., & Rensing, C. (2020) · Lecture notes in computer science · doi:10.1007/978-3-030-52237-7_41 · Cited by 17 · cited 1× · see citing works
“Furthermore, Steuer et al. [25], and Wang et al. [26] evaluated AQG systems trained on SQuAD on educational corpora with promising results.”
BackgroundSelf-citationTowards Automatic Generation of Questions from Long AnswersPreprint
Mishra, S. K., Goel, P., Sharma, A., Jagannatha, A., Jacobs, D. W., & Daumé, H. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2004.05109 · Cited by 11 · cited 1× · see citing works
“Similar research exists for the NQA dataset [9], which is not based on crowd-sourced questions like SQuAD but instead builds on search queries from real users [7].”
Generating Questions for Reading Comprehension using Coherence Relations
Desai, T., Dakle, P. P., & MOLDOVAN, D. I. (2018) · Workshop on Natural Language Processing Techniques for Educational Applications (NLP-TEA) · doi:10.18653/v1/w18-3701 · Cited by 10 · cited 1× · see citing works
“Besides learning the answer selection directly from data, transferring semantically meaningful tagging schemes from other tasks to answer selection has been investigated [31], [32].”
MethodExploring Artificial Jabbering for Automatic Text Comprehension Question Generation
Steuer, T., Filighera, A., & Rensing, C. (2020) · Lecture notes in computer science · doi:10.1007/978-3-030-57717-9_1 · Cited by 8 · see citing works
Self-citationBERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
Devlin, J., Chang, M.-W., Lee, K., & Toutanova, K. (2018) · arXiv · doi:10.48550/arxiv.1810.04805 · cited 2×
“Hence, we deemed six epochs as sufficient for the task, in line with findings by the BERT authors, suggesting that four epochs are sufficient for various fine-tuning tasks [13].”
Method
Flagged first, then by citations
Research works printed but not linked: 1
- Z. Wang, A. S. Lan, W. Nie, A. E. Waters, P. J. Grimaldi, and R. G. Baraniuk, "Qg-net: a data-driven question generation model for educa- tional content," in Proceedings of the Fifth Annual ACM Conference on Learning at Scale, 2018, pp. 1-10.
Cited by
Sorted by citations
Mulla, N., & Gharpure, P. (2023) · Progress in Artificial Intelligence · doi:10.1007/s13748-023-00295-9 · Cited by 132 · see citing works
“Also, most works used factoid-based questions but Yes/No [133], MCQ type [124] and reasoning type questions [122] were also generated for some applications.”
BackgroundMethodInfluentialSurvey of Natural Language Processing for Education: Taxonomy, Systematic Review, and Future Trends
Lan, Y., Li, X., Du, H., Lu, X., Gao, M. Z., Qian, W., & Zhou, A. (2025) · IEEE Transactions on Knowledge and Data Engineering · doi:10.1109/tkde.2025.3621181 · Cited by 43 · see citing works
“The source data contain submissions in programming courses that can be collected on the academy platforms such as JetBrains Academy, MOOC, tutoring sessions and so on [85, 148, 158].”
Pandey, A. K., & Roy, S. S. (2024) · IEEE Access · doi:10.1109/access.2024.3431282 · Cited by 11 · see citing works
“The study also highlights the subtle differences between educational and non-educational AQG tasks [5].”
BackgroundAutomating Turkish Educational Quiz Generation Using Large Language Models
Zeinalipour, K., Keptiğ, Y. G., Maggini, M., & Gori, M. (2025) · Communications in computer and information science · doi:10.1007/978-3-031-82150-9_19 · Cited by 7 · see citing works
Automated Summarization and Question Generation for Efficient Information Extraction
Verma, H., & Anand, N. (2024) · Parul International Conference on Engineering and Technology (PICET) · doi:10.1109/picet60765.2024.10716103 · Cited by 4 · see citing works
Tuhpatussania, S. T., Utami, E., Kusrini, K., & Yuana, K. A. (2024) · International Conference on Technology Innovation and Its Applications (ICTIIA) · doi:10.1109/ictiia61827.2024.10761325 · Cited by 3 · see citing works
“[19], [21], [22], [23], [25], [26], [27], [28], [29], [32], [33], [35], [36], [37], [40], [41], [42], [43], [44], [46], [47], [49], [51], [52], [55] The popular model used today is the Transformer model.”
InfluentialA Review of Large Language Model Evaluation Metrics in Education
Paoprasert, N., Chui, K. T., Yip, T. Y., Liu, J., & Ng, K. (2025) · Communications in computer and information science · doi:10.1007/978-981-95-4499-8_33 · Cited by 3 · see citing works
Diverse Content Selection for Educational Question Generation
Hadifar, A., Bitew, S. K., Deleu, J., Hoste, V., Develder, C., & Demeester, T. (2023) · Conference of the European Chapter of the Association for Computational Linguistics: Student Research Workshop (EACL SRW) · doi:10.18653/v1/2023.eacl-srw.13 · Cited by 2 · see citing works
“As can be seen in the tables, the length of the chapter in TQA-A is significantly shorter than in EduQG.”
BackgroundInfluentialBloom’s Taxonomy based automatic Marathi question generation
(2024) · International Journal of Advanced Technology and Engineering Exploration · doi:10.19101/ijatee.2023.10102163 · Cited by 2 · see citing works
“Steuer et al. [25] investigated the transferability of non-educational answer selection models to the educational domain for AQG.”
Automatic Question Generation for Cognitive Domain of Outcome-Based Education System
Zulfiqar, S., Bazai, S. U., Ghafoor, M. I., Soomro, M., Hussain, L., & Haider, A. (2024) · International Conference on IT and Industrial Technologies (ICIT) · doi:10.1109/icit63607.2024.10860228 · Cited by 1 · see citing works
Attentional Seq2Seq Model for Arabic Opinion Question Generation
Ouahrani, L., & Bennouar, D. (2025) · Lecture notes in networks and systems · doi:10.1007/978-3-031-82112-7_9 · Cited by 1 · see citing works
Research on Question Generation Based on Free-text
Zhou, M., Huang, X., Dong, L., Jiang, X., Yu, L., & Yu, L. (2024) · International Conference on Electronic Information Engineering and Computer Communication (EIECC) · doi:10.1109/eiecc64539.2024.10929146
“QG not only provides datasets for other NLP tasks but also enhances model performance and promotes applications in human-computer interaction and education[2].”
Ganesh, T. N. S., Sheik, R., Subramanyam, S., Thomas, J., & Nirmala, S. J. (2025) · Journal of Ambient Intelligence and Humanized Computing · doi:10.1007/s12652-025-04992-2
“For the Answer Identification module, we used the TQA-A dataset (Steuer et al. 2022), a comprehensive question-answer dataset paired with statements and their corre - sponding answer tags.”
- Authors
- Steuer, T., Filighera, A., & Tregel, T.
- Affiliations
- Technische Universität Darmstadt (DE)
- Venue
- IEEE Access
- Published
- 2022
- Volume
- 10
- Pages
- 63522-63531
- ISSN
- 2169-3536 (ISSN-L)
- OpenAlex
- W4285224010
- Cited by
- 13
- References
- 36
- Citations by year
- 2025: 5 · 2024: 6 · 2023: 2
- FWCI
- 1.07
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- Locations
- IEEE Access (publishedVersion, cc-by-nc-nd)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- Funders
- Bundesministerium für Bildung und Forschung in the Project: Software Campus 2.0 through Microproject: Textbook-Automatic Question Generation (01IS17045)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Topic Modeling
- Physical Sciences › Computer Science › Artificial Intelligence › Natural Language Processing Techniques
- Physical Sciences › Computer Science › Information Systems › Artificial Intelligence in Education
- Keywords
- educational question generation, answer selection, transfer learning, BERT-based models, limited training data
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
