{"url":"/task/text-infilling","name":"Text Infilling","slug":"text-infilling","description_markdown":"**Text Infilling** is the task of predicting missing spans of text which are consistent with the preceding and subsequent text. Text Infilling is a generalization of the cloze task—cloze historically refers to infilling individual words.\n\n\n<span class=\"description-source\">Source: [Enabling Language Models to Fill in the Blanks ](https://arxiv.org/abs/2005.05339)</span>","categories":[{"name":"Adversarial","url":"/area/adversarial"},{"name":"Computer Code","url":"/area/computer-code"},{"name":"Natural Language Processing","url":"/area/natural-language-processing"},{"name":"Speech","url":"/area/speech"}],"source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","slug_source":"archive_url"},"counts":{"papers_tagged":43,"papers_with_code":25,"benchmarks":0,"benchmark_tables_in_archive":0,"benchmark_tables_shown":0,"benchmark_tables_withheld_as_spam":0,"benchmark_definition":"a leaderboard table with at least one row; benchmark_tables_shown also counts the zero-row tables; benchmark_tables_in_archive adds the tables withheld as spam","datasets":1,"subtasks":0,"parent_tasks":1},"benchmarks":[],"datasets":[{"url":"/dataset/wiqa","name":"WIQA","full_name":"What-If Question Answering","num_papers_in_archive":21}],"subtasks":[],"parent_tasks":[{"url":"/task/text-generation","name":"Text Generation"}],"papers":{"order":"repositories listed in the archive (desc), then date (desc); the archive holds no stars","population":"papers tagged with this task that list at least one repository in the archive","shown":25,"of":25,"tagged_in_all":43,"items":[{"url":"/paper/enabling-language-models-to-fill-in-the","title":"Enabling Language Models to Fill in the Blanks","date":"2020-05-11","arxiv_id":"2005.05339","repositories_listed":3,"syntology":{"n":1,"n_ran":1,"n_unverified":0,"n_pointer_only":1}},{"url":"/paper/lot-a-benchmark-for-evaluating-chinese-long","title":"LOT: A Story-Centric Benchmark for Evaluating Chinese Long Text Understanding and Generation","date":"2021-08-30","arxiv_id":"2108.12960","repositories_listed":2,"syntology":null},{"url":"/paper/nutribullets-hybrid-multi-document-health","title":"Nutribullets Hybrid: Multi-document Health Summarization","date":"2021-04-08","arxiv_id":"2104.03465","repositories_listed":2,"syntology":null},{"url":"/paper/the-devil-behind-the-mask-an-emergent-safety","title":"The Devil behind the mask: An emergent safety vulnerability of Diffusion LLMs","date":"2025-07-15","arxiv_id":"2507.11097","repositories_listed":1,"syntology":{"n":17,"n_ran":11,"n_unverified":6,"n_pointer_only":0}},{"url":"/paper/lavida-a-large-diffusion-language-model-for","title":"LaViDa: A Large Diffusion Language Model for Multimodal Understanding","date":"2025-05-22","arxiv_id":"2505.16839","repositories_listed":1,"syntology":{"n":9,"n_ran":3,"n_unverified":6,"n_pointer_only":0}},{"url":"/paper/trajgpt-controlled-synthetic-trajectory","title":"TrajGPT: Controlled Synthetic Trajectory Generation Using a Multitask Transformer-Based Spatiotemporal Model","date":"2024-11-07","arxiv_id":"2411.04381","repositories_listed":1,"syntology":{"n":2,"n_ran":2,"n_unverified":0,"n_pointer_only":2}},{"url":"/paper/empowering-character-level-text-infilling-by","title":"Empowering Character-level Text Infilling by Eliminating Sub-Tokens","date":"2024-05-27","arxiv_id":"2405.17103","repositories_listed":1,"syntology":{"n":1,"n_ran":1,"n_unverified":0,"n_pointer_only":0}},{"url":"/paper/probabilistically-sound-beam-search-with","title":"Towards Probabilistically-Sound Beam Search with Masked Language Models","date":"2024-02-22","arxiv_id":"2402.15020","repositories_listed":1,"syntology":null},{"url":"/paper/a-simple-yet-effective-framework-for-few-shot","title":"A Simple yet Effective Framework for Few-Shot Aspect-Based Sentiment Analysis","date":"2023-07-18","arxiv_id":null,"repositories_listed":1,"syntology":null},{"url":"/paper/having-beer-after-prayer-measuring-cultural","title":"Having Beer after Prayer? Measuring Cultural Bias in Large Language Models","date":"2023-05-23","arxiv_id":"2305.14456","repositories_listed":1,"syntology":null},{"url":"/paper/magvlt-masked-generative-vision-and-language","title":"MAGVLT: Masked Generative Vision-and-Language Transformer","date":"2023-03-21","arxiv_id":"2303.12208","repositories_listed":1,"syntology":{"n":10,"n_ran":7,"n_unverified":3,"n_pointer_only":0}},{"url":"/paper/model-tuning-via-prompts-makes-nlp-models","title":"Model-tuning Via Prompts Makes NLP Models Adversarially Robust","date":"2023-03-13","arxiv_id":"2303.07320","repositories_listed":1,"syntology":null},{"url":"/paper/generative-prompt-tuning-for-relation-1","title":"Generative Prompt Tuning for Relation Classification","date":"2022-10-22","arxiv_id":"2210.12435","repositories_listed":1,"syntology":null},{"url":"/paper/metafill-text-infilling-for-meta-path","title":"MetaFill: Text Infilling for Meta-Path Generation on Heterogeneous Information Networks","date":"2022-10-14","arxiv_id":"2210.07488","repositories_listed":1,"syntology":null},{"url":"/paper/reprogramming-large-pretrained-language","title":"Reprogramming Pretrained Language Models for Antibody Sequence Infilling","date":"2022-10-05","arxiv_id":"2210.07144","repositories_listed":1,"syntology":{"n":7,"n_ran":4,"n_unverified":3,"n_pointer_only":0}},{"url":"/paper/prompting-electra-few-shot-learning-with","title":"Prompting ELECTRA: Few-Shot Learning with Discriminative Pre-Trained Models","date":"2022-05-30","arxiv_id":"2205.15223","repositories_listed":1,"syntology":null},{"url":"/paper/ctrleval-an-unsupervised-reference-free","title":"CTRLEval: An Unsupervised Reference-Free Metric for Evaluating Controlled Text Generation","date":"2022-04-02","arxiv_id":"2204.00862","repositories_listed":1,"syntology":{"n":1,"n_ran":0,"n_unverified":1,"n_pointer_only":1}},{"url":"/paper/language-modeling-via-stochastic-processes-1","title":"Language modeling via stochastic processes","date":"2022-03-21","arxiv_id":"2203.11370","repositories_listed":1,"syntology":null},{"url":"/paper/conformal-prediction-for-text-infilling-and","title":"Conformal prediction for text infilling and part-of-speech prediction","date":"2021-11-04","arxiv_id":"2111.02592","repositories_listed":1,"syntology":null},{"url":"/paper/show-me-how-to-revise-improving-lexically","title":"Show Me How To Revise: Improving Lexically Constrained Sentence Generation with XLNet","date":"2021-09-13","arxiv_id":"2109.05797","repositories_listed":1,"syntology":{"n":1,"n_ran":0,"n_unverified":1,"n_pointer_only":1}},{"url":"/paper/improving-sequence-to-sequence-pre-training","title":"Improving Sequence-to-Sequence Pre-training via Sequence Span Rewriting","date":"2021-01-02","arxiv_id":"2101.00416","repositories_listed":1,"syntology":null},{"url":"/paper/back-to-the-future-unsupervised-backprop","title":"Back to the Future: Unsupervised Backprop-based Decoding for Counterfactual and Abductive Commonsense Reasoning","date":"2020-10-12","arxiv_id":"2010.05906","repositories_listed":1,"syntology":{"n":8,"n_ran":5,"n_unverified":3,"n_pointer_only":8}},{"url":"/paper/keep-calm-and-switch-on-preserving-sentiment","title":"Keep Calm and Switch On! Preserving Sentiment and Fluency in Semantic Text Exchange","date":"2019-08-30","arxiv_id":"1909.00088","repositories_listed":1,"syntology":null},{"url":"/paper/190510752","title":"TIGS: An Inference Algorithm for Text Infilling with Gradient Search","date":"2019-05-26","arxiv_id":"1905.10752","repositories_listed":1,"syntology":null},{"url":"/paper/text-infilling","title":"Text Infilling","date":"2019-01-01","arxiv_id":"1901.00158","repositories_listed":1,"syntology":{"n":4,"n_ran":0,"n_unverified":4,"n_pointer_only":0}}],"syntology_records":11,"syntology_note":"a paper without a record is not a recorded non-run: it may lack an arXiv id or simply be absent from the graph layer"},"description_links":{"kept":0,"unwrapped_to_text":0,"bare_urls_linked":0,"relative_images_dropped":0,"rule":"internal links are kept only when the target slug exists in the catalog"},"syntology":{"read_at":"2026-09-24T18:15:14+00:00","claim":"Per-sample execution status on synthesized fixtures ('ran N of M samples'); not a correctness claim and not a ranking signal.","status_vocabulary":{"ran_honours":"ran, honoured the contract we drafted","ran_violates":"ran, violated the contract we drafted","ran_draft_wrong":"ran; our contract draft was wrong, not the code","ran_fixture":"ran; our fixture could not drive it","ran":"ran on a synthesized input","unverified":"unverified (harvested, no recorded run)"}},"not_shown":{"libraries":"the archive has no per-task library table","trend_sparklines":"the Trend column of the benchmarks table was a rendered image; it is not in the archive","social_and_latest_sorts":"stars and social signals are not in the archive"}}