{"url":"/sota/abstractive-text-summarization-on-cnn-daily","task":{"name":"Abstractive Text Summarization","url":"/task/abstractive-text-summarization","note":null},"dataset":{"name":"CNN / Daily Mail","url":"/dataset/cnn-daily-mail-1"},"category":"Natural Language Processing","categories":["Natural Language Processing"],"category_note":null,"description":"**Abstractive Text Summarization** is the task of generating a short and concise summary that captures the salient ideas of the source text. The generated summaries potentially contain new phrases and sentences that may not appear in the source text.\r\n\r\n\r\n<span class=\"description-source\">Source: [Generative Adversarial Network for Abstractive Text Summarization ](https://arxiv.org/abs/1711.09357)</span>\r\n\r\nImage credit: [Abstractive Text Summarization using Sequence-to-sequence RNNs and Beyond](https://arxiv.org/pdf/1602.06023v5.pdf)","description_from":"task","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","rank":"the archive's row order at snapshot; not re-ranked","rows_end_at":"2025-07-28","rows_withheld_as_spam":0,"metric_values":"the archive's strings, untouched"},"metrics":["ROUGE-1","ROUGE-2","ROUGE-L"],"metric_direction":{"note":"inferred from the metric name only (the archive records no direction); null = not inferred, chart draws points only","by_metric":{"ROUGE-1":"higher","ROUGE-2":"higher","ROUGE-L":"higher"}},"counts":{"rows":53,"rows_with_code":42,"rows_with_paper_page":53,"rows_dated":51,"rows_using_additional_data":11},"rows":[{"rank_in_archive_order":1,"model":"Scrambled code + broken (alter)","metrics":{"ROUGE-1":"48.18","ROUGE-2":"19.84","ROUGE-L":"45.35"},"uses_additional_data":false,"paper_date":"2022-10-25","paper":"/paper/universal-evasion-attacks-on-summarization","paper_url":"https://arxiv.org/abs/2210.14260v1","paper_title":"Universal Evasion Attacks on Summarization Scoring","code":"https://github.com/cestwc/universal-evasion","n_code_links":1,"syntology":null},{"rank_in_archive_order":2,"model":"BRIO","metrics":{"ROUGE-1":"47.78","ROUGE-2":"23.55","ROUGE-L":"44.57"},"uses_additional_data":false,"paper_date":"2022-03-31","paper":"/paper/brio-bringing-order-to-abstractive","paper_url":"https://arxiv.org/abs/2203.16804v1","paper_title":"BRIO: Bringing Order to Abstractive Summarization","code":"https://github.com/yixinl7/brio","n_code_links":3,"syntology":{"n_ran":7,"n_unverified":3,"n_samples":10,"n_pointer_only_licence":0}},{"rank_in_archive_order":3,"model":"Pegasus","metrics":{"ROUGE-1":"47.36","ROUGE-2":"24.02","ROUGE-L":"44.45"},"uses_additional_data":false,"paper_date":"2022-09-30","paper":"/paper/calibrating-sequence-likelihood-improves","paper_url":"https://arxiv.org/abs/2210.00045v1","paper_title":"Calibrating Sequence likelihood Improves Conditional Language Generation","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":4,"model":"PEGASUS + SummaReranker","metrics":{"ROUGE-1":"47.16","ROUGE-2":"22.61","ROUGE-L":"43.87"},"uses_additional_data":false,"paper_date":"2022-03-13","paper":"/paper/summareranker-a-multi-task-mixture-of-experts-1","paper_url":"https://arxiv.org/abs/2203.06569v2","paper_title":"SummaReranker: A Multi-Task Mixture-of-Experts Re-ranking Framework for Abstractive Summarization","code":"https://github.com/ntunlp/summareranker","n_code_links":1,"syntology":{"n_ran":2,"n_unverified":1,"n_samples":3,"n_pointer_only_licence":0}},{"rank_in_archive_order":5,"model":"Scrambled code + broken","metrics":{"ROUGE-1":"46.71","ROUGE-2":"20.39","ROUGE-L":"43.56"},"uses_additional_data":false,"paper_date":"2022-10-25","paper":"/paper/universal-evasion-attacks-on-summarization","paper_url":"https://arxiv.org/abs/2210.14260v1","paper_title":"Universal Evasion Attacks on Summarization Scoring","code":"https://github.com/cestwc/universal-evasion","n_code_links":1,"syntology":null},{"rank_in_archive_order":6,"model":"BART + SimCLS","metrics":{"ROUGE-1":"46.67","ROUGE-2":"22.15","ROUGE-L":"43.54"},"uses_additional_data":false,"paper_date":"2021-06-03","paper":"/paper/simcls-a-simple-framework-for-contrastive","paper_url":"https://arxiv.org/abs/2106.01890v1","paper_title":"SimCLS: A Simple Framework for Contrastive Learning of Abstractive Summarization","code":"https://github.com/yixinL7/SimCLS","n_code_links":2,"syntology":{"n_ran":4,"n_unverified":2,"n_samples":6,"n_pointer_only_licence":1}},{"rank_in_archive_order":7,"model":"SEASON","metrics":{"ROUGE-1":"46.27","ROUGE-2":"22.64","ROUGE-L":"43.08"},"uses_additional_data":false,"paper_date":"2022-10-22","paper":"/paper/salience-allocation-as-guidance-for","paper_url":"https://arxiv.org/abs/2210.12330v1","paper_title":"Salience Allocation as Guidance for Abstractive Summarization","code":"https://github.com/tencent-ailab/season","n_code_links":1,"syntology":null},{"rank_in_archive_order":8,"model":"Fourier Transformer","metrics":{"ROUGE-1":"44.76","ROUGE-2":"21.55","ROUGE-L":"41.34"},"uses_additional_data":false,"paper_date":"2023-05-24","paper":"/paper/fourier-transformer-fast-long-range-modeling","paper_url":"https://arxiv.org/abs/2305.15099v2","paper_title":"Fourier Transformer: Fast Long Range Modeling by Removing Sequence Redundancy with FFT Operator","code":"https://github.com/lumia-group/fouriertransformer","n_code_links":1,"syntology":null},{"rank_in_archive_order":9,"model":"GLM-XXLarge","metrics":{"ROUGE-1":"44.7","ROUGE-2":"21.4","ROUGE-L":"41.4"},"uses_additional_data":true,"paper_date":"2021-03-18","paper":"/paper/all-nlp-tasks-are-generation-tasks-a-general","paper_url":"https://arxiv.org/abs/2103.10360v2","paper_title":"GLM: General Language Model Pretraining with Autoregressive Blank Infilling","code":"https://github.com/thudm/chatglm2-6b","n_code_links":8,"syntology":{"n_ran":1,"n_unverified":0,"n_samples":1,"n_pointer_only_licence":0}},{"rank_in_archive_order":10,"model":"BART + R-Drop","metrics":{"ROUGE-1":"44.51","ROUGE-2":"21.58","ROUGE-L":"41.24"},"uses_additional_data":false,"paper_date":"2021-06-28","paper":"/paper/r-drop-regularized-dropout-for-neural","paper_url":"https://arxiv.org/abs/2106.14448v2","paper_title":"R-Drop: Regularized Dropout for Neural Networks","code":"https://github.com/dropreg/R-Drop","n_code_links":8,"syntology":{"n_ran":4,"n_unverified":2,"n_samples":6,"n_pointer_only_licence":0}},{"rank_in_archive_order":11,"model":"CoCoNet + CoCoPretrain","metrics":{"ROUGE-1":"44.50","ROUGE-2":"21.55","ROUGE-L":"41.24"},"uses_additional_data":true,"paper_date":null,"paper":"/paper/learn-to-copy-from-the-copying-history","paper_url":"https://aclanthology.org/2021.emnlp-main.336","paper_title":"Learn to Copy from the Copying History: Correlational Copy Network for Abstractive Summarization","code":"https://github.com/hrlinlp/coconet","n_code_links":1,"syntology":null},{"rank_in_archive_order":12,"model":"MUPPET BART Large","metrics":{"ROUGE-1":"44.45","ROUGE-2":"21.25","ROUGE-L":"41.4"},"uses_additional_data":false,"paper_date":"2021-01-26","paper":"/paper/muppet-massive-multi-task-representations","paper_url":"https://arxiv.org/abs/2101.11038v1","paper_title":"Muppet: Massive Multi-task Representations with Pre-Finetuning","code":"https://huggingface.co/facebook/muppet-roberta-base","n_code_links":2,"syntology":null},{"rank_in_archive_order":13,"model":"CoCoNet","metrics":{"ROUGE-1":"44.39","ROUGE-2":"21.41","ROUGE-L":"41.05"},"uses_additional_data":false,"paper_date":null,"paper":"/paper/learn-to-copy-from-the-copying-history","paper_url":"https://aclanthology.org/2021.emnlp-main.336","paper_title":"Learn to Copy from the Copying History: Correlational Copy Network for Abstractive Summarization","code":"https://github.com/hrlinlp/coconet","n_code_links":1,"syntology":null},{"rank_in_archive_order":14,"model":"BART+R3F","metrics":{"ROUGE-1":"44.38","ROUGE-2":"21.53","ROUGE-L":"41.17"},"uses_additional_data":false,"paper_date":"2020-08-06","paper":"/paper/better-fine-tuning-by-reducing","paper_url":"https://arxiv.org/abs/2008.03156v1","paper_title":"Better Fine-Tuning by Reducing Representational Collapse","code":"https://github.com/pytorch/fairseq/tree/master/examples/rxf","n_code_links":3,"syntology":null},{"rank_in_archive_order":15,"model":"ERNIE-GENLARGE (large-scale text corpora)","metrics":{"ROUGE-1":"44.31","ROUGE-2":"21.35","ROUGE-L":"41.60"},"uses_additional_data":true,"paper_date":"2020-01-26","paper":"/paper/ernie-gen-an-enhanced-multi-flow-pre-training","paper_url":"https://arxiv.org/abs/2001.11314v3","paper_title":"ERNIE-GEN: An Enhanced Multi-Flow Pre-training and Fine-tuning Framework for Natural Language Generation","code":"https://github.com/PaddlePaddle/PaddleNLP/blob/develop/paddlenlp/transformers/ernie_gen/modeling.py","n_code_links":5,"syntology":null},{"rank_in_archive_order":16,"model":"PALM","metrics":{"ROUGE-1":"44.30","ROUGE-2":"21.12","ROUGE-L":"41.41"},"uses_additional_data":false,"paper_date":"2020-04-14","paper":"/paper/palm-pre-training-an-autoencoding","paper_url":"https://arxiv.org/abs/2004.07159v2","paper_title":"PALM: Pre-training an Autoencoding&Autoregressive Language Model for Context-conditioned Generation","code":"https://github.com/alibaba/AliceMind/tree/main/PALM","n_code_links":2,"syntology":null},{"rank_in_archive_order":17,"model":"ProphetNet","metrics":{"ROUGE-1":"44.20","ROUGE-2":"21.17","ROUGE-L":"41.30"},"uses_additional_data":true,"paper_date":"2020-01-13","paper":"/paper/prophetnet-predicting-future-n-gram-for","paper_url":"https://arxiv.org/abs/2001.04063v3","paper_title":"ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training","code":"https://github.com/huggingface/transformers","n_code_links":5,"syntology":null},{"rank_in_archive_order":18,"model":"PEGASUS","metrics":{"ROUGE-1":"44.17","ROUGE-2":"21.47","ROUGE-L":"41.11"},"uses_additional_data":true,"paper_date":"2019-12-18","paper":"/paper/pegasus-pre-training-with-extracted-gap","paper_url":"https://arxiv.org/abs/1912.08777v2","paper_title":"PEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarization","code":"https://github.com/huggingface/transformers","n_code_links":19,"syntology":{"n_ran":15,"n_unverified":4,"n_samples":19,"n_pointer_only_licence":0}},{"rank_in_archive_order":19,"model":"BART","metrics":{"ROUGE-1":"44.16","ROUGE-2":"21.28","ROUGE-L":"40.90"},"uses_additional_data":false,"paper_date":"2019-10-29","paper":"/paper/bart-denoising-sequence-to-sequence-pre","paper_url":"https://arxiv.org/abs/1910.13461v1","paper_title":"BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehension","code":"https://github.com/huggingface/transformers","n_code_links":47,"syntology":{"n_ran":39,"n_unverified":14,"n_samples":53,"n_pointer_only_licence":7}},{"rank_in_archive_order":20,"model":"ERNIE-GENLARGE","metrics":{"ROUGE-1":"44.02","ROUGE-2":"21.17","ROUGE-L":"41.26"},"uses_additional_data":false,"paper_date":"2020-01-26","paper":"/paper/ernie-gen-an-enhanced-multi-flow-pre-training","paper_url":"https://arxiv.org/abs/2001.11314v3","paper_title":"ERNIE-GEN: An Enhanced Multi-Flow Pre-training and Fine-tuning Framework for Natural Language Generation","code":"https://github.com/PaddlePaddle/PaddleNLP/blob/develop/paddlenlp/transformers/ernie_gen/modeling.py","n_code_links":5,"syntology":null},{"rank_in_archive_order":21,"model":"LongT5","metrics":{"ROUGE-1":"43.94","ROUGE-2":"21.40 ","ROUGE-L":"41.28"},"uses_additional_data":false,"paper_date":"2021-12-15","paper":"/paper/longt5-efficient-text-to-text-transformer-for","paper_url":"https://arxiv.org/abs/2112.07916v2","paper_title":"LongT5: Efficient Text-To-Text Transformer for Long Sequences","code":"https://github.com/google-research/longt5","n_code_links":4,"syntology":{"n_ran":1,"n_unverified":0,"n_samples":1,"n_pointer_only_licence":1}},{"rank_in_archive_order":22,"model":"T5","metrics":{"ROUGE-1":"43.52","ROUGE-2":"21.55","ROUGE-L":"40.69"},"uses_additional_data":false,"paper_date":"2019-10-23","paper":"/paper/exploring-the-limits-of-transfer-learning","paper_url":"https://arxiv.org/abs/1910.10683v4","paper_title":"Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer","code":"https://github.com/huggingface/transformers","n_code_links":57,"syntology":{"n_ran":21,"n_unverified":10,"n_samples":31,"n_pointer_only_licence":0}},{"rank_in_archive_order":23,"model":"SRformer-BART","metrics":{"ROUGE-1":"43.19","ROUGE-2":"19.80","ROUGE-L":"40.40"},"uses_additional_data":false,"paper_date":"2023-05-24","paper":"/paper/segmented-recurrent-transformer-an-efficient","paper_url":"https://arxiv.org/abs/2305.16340v3","paper_title":"Segmented Recurrent Transformer: An Efficient Sequence-to-Sequence Model","code":"https://github.com/yinghanlong/SRtransformer","n_code_links":1,"syntology":null},{"rank_in_archive_order":24,"model":"UniLMv2","metrics":{"ROUGE-1":"43.16","ROUGE-2":"20.42","ROUGE-L":"40.14"},"uses_additional_data":true,"paper_date":"2020-02-28","paper":"/paper/unilmv2-pseudo-masked-language-models-for","paper_url":"https://arxiv.org/abs/2002.12804v1","paper_title":"UniLMv2: Pseudo-Masked Language Models for Unified Language Model Pre-Training","code":"https://github.com/microsoft/unilm","n_code_links":3,"syntology":{"n_ran":1,"n_unverified":0,"n_samples":1,"n_pointer_only_licence":0}},{"rank_in_archive_order":25,"model":"UniLM","metrics":{"ROUGE-1":"43.08","ROUGE-2":"20.43","ROUGE-L":"40.34"},"uses_additional_data":true,"paper_date":"2019-05-08","paper":"/paper/unified-language-model-pre-training-for","paper_url":"https://arxiv.org/abs/1905.03197v3","paper_title":"Unified Language Model Pre-training for Natural Language Understanding and Generation","code":"https://github.com/microsoft/unilm","n_code_links":9,"syntology":null},{"rank_in_archive_order":26,"model":"ERNIE-GENBASE","metrics":{"ROUGE-1":"42.30","ROUGE-2":"19.92","ROUGE-L":"39.68"},"uses_additional_data":true,"paper_date":"2020-01-26","paper":"/paper/ernie-gen-an-enhanced-multi-flow-pre-training","paper_url":"https://arxiv.org/abs/2001.11314v3","paper_title":"ERNIE-GEN: An Enhanced Multi-Flow Pre-training and Fine-tuning Framework for Natural Language Generation","code":"https://github.com/PaddlePaddle/PaddleNLP/blob/develop/paddlenlp/transformers/ernie_gen/modeling.py","n_code_links":5,"syntology":null},{"rank_in_archive_order":27,"model":"BertSumExtAbs","metrics":{"ROUGE-1":"42.13","ROUGE-2":"19.6","ROUGE-L":"39.18"},"uses_additional_data":true,"paper_date":"2019-08-22","paper":"/paper/text-summarization-with-pretrained-encoders","paper_url":"https://arxiv.org/abs/1908.08345v2","paper_title":"Text Summarization with Pretrained Encoders","code":"https://github.com/nlpyang/PreSumm","n_code_links":19,"syntology":{"n_ran":11,"n_unverified":10,"n_samples":21,"n_pointer_only_licence":0}},{"rank_in_archive_order":28,"model":"BERT-ext + abs + RL + rerank","metrics":{"ROUGE-1":"41.90","ROUGE-2":"19.08","ROUGE-L":"39.64"},"uses_additional_data":false,"paper_date":"2019-09-19","paper":"/paper/summary-level-training-of-sentence-rewriting","paper_url":"https://arxiv.org/abs/1909.08752v3","paper_title":"Summary Level Training of Sentence Rewriting for Abstractive Summarization","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":29,"model":"Selector & Pointer-Generator","metrics":{"ROUGE-1":"41.72","ROUGE-2":"18.74","ROUGE-L":"38.79"},"uses_additional_data":true,"paper_date":"2019-09-04","paper":"/paper/mixture-content-selection-for-diverse","paper_url":"https://arxiv.org/abs/1909.01953v1","paper_title":"Mixture Content Selection for Diverse Sequence Generation","code":"https://github.com/clovaai/FocusSeq2Seq","n_code_links":1,"syntology":null},{"rank_in_archive_order":30,"model":"Two-Stage + RL","metrics":{"ROUGE-1":"41.71","ROUGE-2":"19.49","ROUGE-L":"38.79"},"uses_additional_data":false,"paper_date":"2019-02-25","paper":"/paper/pretraining-based-natural-language-generation","paper_url":"http://arxiv.org/abs/1902.09243v2","paper_title":"Pretraining-Based Natural Language Generation for Text Summarization","code":"https://github.com/nayeon7lee/bert-summarization","n_code_links":4,"syntology":{"n_ran":1,"n_unverified":0,"n_samples":1,"n_pointer_only_licence":1}},{"rank_in_archive_order":31,"model":"DCA","metrics":{"ROUGE-1":"41.69","ROUGE-2":"19.47","ROUGE-L":"37.92"},"uses_additional_data":true,"paper_date":"2018-03-27","paper":"/paper/deep-communicating-agents-for-abstractive","paper_url":"http://arxiv.org/abs/1803.10357v3","paper_title":"Deep Communicating Agents for Abstractive Summarization","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":32,"model":"Li et al.","metrics":{"ROUGE-1":"41.54","ROUGE-2":"18.18","ROUGE-L":"36.47"},"uses_additional_data":false,"paper_date":"2018-10-01","paper":"/paper/improving-neural-abstractive-document","paper_url":"https://aclanthology.org/D18-1205","paper_title":"Improving Neural Abstractive Document Summarization with Explicit Information Selection Modeling","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":33,"model":"rnn-ext + RL","metrics":{"ROUGE-1":"41.47","ROUGE-2":"18.72","ROUGE-L":"37.76"},"uses_additional_data":false,"paper_date":"2018-05-28","paper":"/paper/fast-abstractive-summarization-with-reinforce","paper_url":"http://arxiv.org/abs/1805.11080v1","paper_title":"Fast Abstractive Summarization with Reinforce-Selected Sentence Rewriting","code":"https://github.com/ChenRocks/fast_abs_rl","n_code_links":3,"syntology":{"n_ran":3,"n_unverified":0,"n_samples":3,"n_pointer_only_licence":2}},{"rank_in_archive_order":34,"model":"EditNet","metrics":{"ROUGE-1":"41.42","ROUGE-2":"19.03","ROUGE-L":"38.36"},"uses_additional_data":false,"paper_date":"2019-02-27","paper":"/paper/an-editorial-network-for-enhanced-document","paper_url":"http://arxiv.org/abs/1902.10360v1","paper_title":"An Editorial Network for Enhanced Document Summarization","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":35,"model":"Bottom-Up Summarization","metrics":{"ROUGE-1":"41.22","ROUGE-2":"18.68","ROUGE-L":"38.34"},"uses_additional_data":false,"paper_date":"2018-08-31","paper":"/paper/bottom-up-abstractive-summarization","paper_url":"http://arxiv.org/abs/1808.10792v2","paper_title":"Bottom-Up Abstractive Summarization","code":"https://github.com/sebastianGehrmann/bottom-up-summary","n_code_links":5,"syntology":{"n_ran":5,"n_unverified":1,"n_samples":6,"n_pointer_only_licence":0}},{"rank_in_archive_order":36,"model":"Mask Attention Network","metrics":{"ROUGE-1":"40.98","ROUGE-2":"18.29","ROUGE-L":"37.88"},"uses_additional_data":false,"paper_date":"2021-03-25","paper":"/paper/mask-attention-networks-rethinking-and","paper_url":"https://arxiv.org/abs/2103.13597v1","paper_title":"Mask Attention Networks: Rethinking and Strengthen Transformer","code":"https://github.com/libertfan/man","n_code_links":1,"syntology":{"n_ran":0,"n_unverified":3,"n_samples":3,"n_pointer_only_licence":0}},{"rank_in_archive_order":37,"model":"Subformer-base","metrics":{"ROUGE-1":"40.9","ROUGE-2":"18.3","ROUGE-L":"37.7"},"uses_additional_data":false,"paper_date":"2021-01-01","paper":"/paper/subformer-a-parameter-reduced-transformer","paper_url":"https://openreview.net/forum?id=6UurSaf08jx","paper_title":"Subformer: A Parameter Reduced Transformer","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":38,"model":"rnn-ext + abs + RL + rerank","metrics":{"ROUGE-1":"40.88","ROUGE-2":"17.80","ROUGE-L":"38.54"},"uses_additional_data":false,"paper_date":"2018-05-28","paper":"/paper/fast-abstractive-summarization-with-reinforce","paper_url":"http://arxiv.org/abs/1805.11080v1","paper_title":"Fast Abstractive Summarization with Reinforce-Selected Sentence Rewriting","code":"https://github.com/ChenRocks/fast_abs_rl","n_code_links":3,"syntology":{"n_ran":3,"n_unverified":0,"n_samples":3,"n_pointer_only_licence":2}},{"rank_in_archive_order":39,"model":"end2end w/ inconsistency loss","metrics":{"ROUGE-1":"40.68","ROUGE-2":"17.97","ROUGE-L":"37.13"},"uses_additional_data":false,"paper_date":"2018-05-16","paper":"/paper/a-unified-model-for-extractive-and","paper_url":"http://arxiv.org/abs/1805.06266v2","paper_title":"A Unified Model for Extractive and Abstractive Summarization using Inconsistency Loss","code":"https://github.com/HsuWanTing/unified-summarization","n_code_links":1,"syntology":{"n_ran":1,"n_unverified":6,"n_samples":7,"n_pointer_only_licence":0}},{"rank_in_archive_order":40,"model":"RL + pg + cbdec","metrics":{"ROUGE-1":"40.66","ROUGE-2":"17.87","ROUGE-L":"37.06"},"uses_additional_data":false,"paper_date":"2018-09-12","paper":"/paper/closed-book-training-to-improve-summarization","paper_url":"http://arxiv.org/abs/1809.04585v1","paper_title":"Closed-Book Training to Improve Summarization Encoder Memory","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":41,"model":"ROUGESal+Ent RL","metrics":{"ROUGE-1":"40.43","ROUGE-2":"18.00","ROUGE-L":"37.10"},"uses_additional_data":false,"paper_date":"2018-04-17","paper":"/paper/multi-reward-reinforced-summarization-with","paper_url":"http://arxiv.org/abs/1804.06451v2","paper_title":"Multi-Reward Reinforced Summarization with Saliency and Entailment","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":42,"model":"LEAD-3","metrics":{"ROUGE-1":"40.42","ROUGE-2":"17.62","ROUGE-L":"36.67"},"uses_additional_data":false,"paper_date":"2016-02-19","paper":"/paper/abstractive-text-summarization-using-sequence","paper_url":"http://arxiv.org/abs/1602.06023v5","paper_title":"Abstractive Text Summarization Using Sequence-to-Sequence RNNs and Beyond","code":"https://github.com/theamrzaki/text_summurization_abstractive_methods","n_code_links":4,"syntology":{"n_ran":5,"n_unverified":1,"n_samples":6,"n_pointer_only_licence":0}},{"rank_in_archive_order":43,"model":"Li et al.","metrics":{"ROUGE-1":"40.30","ROUGE-2":"18.02","ROUGE-L":"37.36"},"uses_additional_data":false,"paper_date":"2018-10-01","paper":"/paper/improving-neural-abstractive-document-1","paper_url":"https://aclanthology.org/D18-1441","paper_title":"Improving Neural Abstractive Document Summarization with Structural Regularization","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":44,"model":"ML+RL ROUGE+Novel, with LM","metrics":{"ROUGE-1":"40.19","ROUGE-2":"17.38","ROUGE-L":"37.52"},"uses_additional_data":false,"paper_date":"2018-08-23","paper":"/paper/improving-abstraction-in-text-summarization","paper_url":"http://arxiv.org/abs/1808.07913v1","paper_title":"Improving Abstraction in Text Summarization","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":45,"model":"Dynamic Conv","metrics":{"ROUGE-1":"39.84","ROUGE-2":"16.25","ROUGE-L":"36.73"},"uses_additional_data":false,"paper_date":"2019-01-29","paper":"/paper/pay-less-attention-with-lightweight-and","paper_url":"http://arxiv.org/abs/1901.10430v2","paper_title":"Pay Less Attention with Lightweight and Dynamic Convolutions","code":"https://github.com/pytorch/fairseq","n_code_links":3,"syntology":null},{"rank_in_archive_order":46,"model":"Pointer + Coverage + EntailmentGen + QuestionGen","metrics":{"ROUGE-1":"39.81","ROUGE-2":"17.64","ROUGE-L":"36.54"},"uses_additional_data":false,"paper_date":"2018-05-28","paper":"/paper/soft-layer-specific-multi-task-summarization","paper_url":"http://arxiv.org/abs/1805.11004v1","paper_title":"Soft Layer-Specific Multi-Task Summarization with Entailment and Question Generation","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":47,"model":"PTGEN + Coverage","metrics":{"ROUGE-1":"39.53","ROUGE-2":"17.28","ROUGE-L":"36.38"},"uses_additional_data":false,"paper_date":"2017-04-14","paper":"/paper/get-to-the-point-summarization-with-pointer","paper_url":"http://arxiv.org/abs/1704.04368v2","paper_title":"Get To The Point: Summarization with Pointer-Generator Networks","code":"https://github.com/PaddlePaddle/PaddleNLP/tree/develop/examples/text_summarization/pointer_summarizer","n_code_links":39,"syntology":{"n_ran":30,"n_unverified":34,"n_samples":64,"n_pointer_only_licence":44}},{"rank_in_archive_order":48,"model":"PTGEN + Coverage","metrics":{"ROUGE-1":"39.53","ROUGE-2":"17.28","ROUGE-L":"36.38"},"uses_additional_data":false,"paper_date":"2017-04-14","paper":"/paper/get-to-the-point-summarization-with-pointer","paper_url":"http://arxiv.org/abs/1704.04368v2","paper_title":"Get To The Point: Summarization with Pointer-Generator Networks","code":"https://github.com/PaddlePaddle/PaddleNLP/tree/develop/examples/text_summarization/pointer_summarizer","n_code_links":39,"syntology":{"n_ran":30,"n_unverified":34,"n_samples":64,"n_pointer_only_licence":44}},{"rank_in_archive_order":49,"model":"Pointer-Generator + Coverage","metrics":{"ROUGE-1":"39.53","ROUGE-2":"17.28"},"uses_additional_data":false,"paper_date":"2017-04-14","paper":"/paper/get-to-the-point-summarization-with-pointer","paper_url":"http://arxiv.org/abs/1704.04368v2","paper_title":"Get To The Point: Summarization with Pointer-Generator Networks","code":"https://github.com/PaddlePaddle/PaddleNLP/tree/develop/examples/text_summarization/pointer_summarizer","n_code_links":39,"syntology":{"n_ran":30,"n_unverified":34,"n_samples":64,"n_pointer_only_licence":44}},{"rank_in_archive_order":50,"model":"Transformer","metrics":{"ROUGE-1":"39.50","ROUGE-2":"16.06","ROUGE-L":"36.63"},"uses_additional_data":false,"paper_date":"2017-06-12","paper":"/paper/attention-is-all-you-need","paper_url":"https://arxiv.org/abs/1706.03762v7","paper_title":"Attention Is All You Need","code":"https://github.com/huggingface/transformers","n_code_links":595,"syntology":{"n_ran":610,"n_unverified":336,"n_samples":946,"n_pointer_only_licence":451}},{"rank_in_archive_order":51,"model":"Summary Loop Unsup","metrics":{"ROUGE-1":"37.7"},"uses_additional_data":false,"paper_date":"2021-05-11","paper":"/paper/the-summary-loop-learning-to-write-1","paper_url":"https://arxiv.org/abs/2105.05361v1","paper_title":"The Summary Loop: Learning to Write Abstractive Summaries Without Examples","code":"https://github.com/cannylab/summary_loop","n_code_links":1,"syntology":null},{"rank_in_archive_order":52,"model":"CriSPO 3-shot","metrics":{"ROUGE-L":"27.4"},"uses_additional_data":false,"paper_date":"2024-10-03","paper":"/paper/crispo-multi-aspect-critique-suggestion","paper_url":"https://arxiv.org/abs/2410.02748v3","paper_title":"CriSPO: Multi-Aspect Critique-Suggestion-guided Automatic Prompt Optimization for Text Generation","code":"https://github.com/amazon-science/crispo","n_code_links":1,"syntology":null},{"rank_in_archive_order":53,"model":"DELTA (BLSTM)","metrics":{"ROUGE-L":"27.3"},"uses_additional_data":false,"paper_date":"2019-08-02","paper":"/paper/delta-a-deep-learning-based-language","paper_url":"https://arxiv.org/abs/1908.01853v1","paper_title":"DELTA: A DEep learning based Language Technology plAtform","code":"https://github.com/didi/delta","n_code_links":2,"syntology":null}],"since_archive":{"claim":"Results that newer papers report for their own method, placed here by Syntology. A model pointed at the cell in the paper's own table; the number was read from that cell and checked against this leaderboard's metric, dataset, split and scale; an independent check that saw this leaderboard's other rows and every other leaderboard on the same dataset accepted it. Not reviewed by the paper's authors or by the archive's editors, and not ranked against the archive rows.","extraction_file_present":true,"measurement":{"test_papers":883,"papers_with_output":881,"judged_true":108,"judged":110,"wilson95_lower":0.9361,"measured_on":"2026-09-24","frozen_commit":"0e3de0df94"},"measurement_note":"blind adjudication of accepted entries on a held-out split of archive papers, rules frozen before the test","coverage":{"sentence":"Syntology has checked 7,081 of the 9,623 papers on this site that are newer than the archive; results from the others appear after they are checked.","complete":false,"papers_newer_than_archive":9623,"papers_checked":7081,"papers_extracted_not_yet_verified":217,"boards_without_verdict":27,"papers_not_yet_extracted":2325},"order":"newest first by month (arXiv date, else the arXiv-id month), then arXiv id descending","columns":[],"entries":[]},"syntology":{"read_at":"2026-09-25T09:33:49+00:00","claim":"Per row: N of M harvested code samples from that row's paper executed on a synthesized fixture; the other M-N are unverified. Not a reproduction of the row's number; not a correctness claim. n_pointer_only_licence counts samples the site points at rather than redistributes (a licence axis, independent of ran/unverified).","rows_with_graph_line":22,"rows_with_any_sample_ran":21,"distinct_papers_with_graph_line":19,"distinct_papers_with_any_sample_ran":18,"samples_over_distinct_papers":{"n_ran":761,"n_unverified":427,"n_samples":1188,"n_pointer_only_licence":507,"note":"each paper (arXiv id) counted once, however many rows it is behind; this is the page-level figure"},"samples_row_weighted":{"n_ran":824,"n_unverified":495,"n_samples":1319,"n_pointer_only_licence":597,"note":"row-weighted: a paper behind several rows is counted once per row; inflated relative to samples_over_distinct_papers by design, kept for readers summing the per-row syntology blocks"}}}