{"id":48978303,"url":"https://github.com/jpwahle/emnlp22-transforming","last_synced_at":"2026-04-18T10:21:38.393Z","repository":{"id":74275020,"uuid":"547299920","full_name":"jpwahle/emnlp22-transforming","owner":"jpwahle","description":"The official implementation of the EMNLP 2022 paper \"How Large Language Models are Transforming Machine-Paraphrased Plagiarism\".","archived":false,"fork":false,"pushed_at":"2023-12-20T21:04:15.000Z","size":799,"stargazers_count":10,"open_issues_count":8,"forks_count":0,"subscribers_count":3,"default_branch":"main","last_synced_at":"2025-09-05T11:12:10.567Z","etag":null,"topics":["machine-learning","natural-language-processing","nlp","paraphrase-generation","plagiarism"],"latest_commit_sha":null,"homepage":"https://aclanthology.org/2022.emnlp-main.62/","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/jpwahle.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"code_of_conduct":null,"threat_model":null,"audit":null,"citation":"CITATION.bib","codeowners":null,"security":null,"support":null,"governance":null}},"created_at":"2022-10-07T13:06:06.000Z","updated_at":"2024-02-08T04:22:18.000Z","dependencies_parsed_at":"2023-10-03T23:59:00.038Z","dependency_job_id":null,"html_url":"https://github.com/jpwahle/emnlp22-transforming","commit_stats":null,"previous_names":[],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/jpwahle/emnlp22-transforming","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/jpwahle%2Femnlp22-transforming","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/jpwahle%2Femnlp22-transforming/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/jpwahle%2Femnlp22-transforming/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/jpwahle%2Femnlp22-transforming/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/jpwahle","download_url":"https://codeload.github.com/jpwahle/emnlp22-transforming/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/jpwahle%2Femnlp22-transforming/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":31964970,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-04-18T00:39:45.007Z","status":"online","status_checked_at":"2026-04-18T02:00:07.018Z","response_time":103,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["machine-learning","natural-language-processing","nlp","paraphrase-generation","plagiarism"],"created_at":"2026-04-18T10:21:37.709Z","updated_at":"2026-04-18T10:21:38.374Z","avatar_url":"https://github.com/jpwahle.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# How Large Language Models Are Transforming Machine Paraphrase Generation\n\n[![arXiv](https://img.shields.io/badge/arXiv-2210.03568-b31b1b.svg)](https://arxiv.org/abs/2210.03568)\n[![HuggingFace Dataset](https://img.shields.io/badge/🤗-Datasets-ffce1c.svg)](https://huggingface.co/datasets/jpwahle/autoregressive-paraphrase-dataset)\n\n## Quick Start\n\n### Install\n\n```bash\npoetry install\n```\n\n### Run\n\nTo generate paraphrases using T5, run the following command:\n\n\u003e Note: T5 benefits from more few shot examples as it actually performs some gradient steps. However, to make it comparable to GPT-3, we don't recommend exceeding 50 examples.\n\n```bash\npoetry run python paraphrase.generate --model_name gpt3 --num_prompts 4 --num_examples 32\n```\n\nFor generating paraphrases using GPT-3, run the following command:\n\n\u003e Warning: Using GPT-3 requires a paid account and can quickly run up a bill if you don't have credits.\n\u003e Reducing the number of prompts and/or the number of samples can help reduce costs.\n\n```bash\nOPENAI_API_KEY={YOUR_KEY} poetry run python paraphrase.generate --model_name gpt3 --num_prompts 4 --num_examples 32\n```\n\nFor help, run the following command:\n\n```bash\npoetry run python -m paraphrase.generate --help\n```\n\n## Dataset\n\nThe dataset generated for our study is available on [🤗 Hugging Face Datasets](https://huggingface.co/datasets/jpwahle/autoregressive-paraphrase-dataset).\n\n## Detection\n\nFor the detection code, please refer to this [repository](https://github.com/jpwahle/iconf22-paraphrase) and [paper](https://link.springer.com/chapter/10.1007/978-3-030-96957-8_34).\n\nFor all models except GPT-3 and T5, we used the trained versions on MPC. For PlagScan, we embedded the text in the same way as in the paper above.\n \n## Citation\n```bib\n@inproceedings{wahle-etal-2022-large,\n    title = \"How Large Language Models are Transforming Machine-Paraphrase Plagiarism\",\n    author = \"Wahle, Jan Philip  and\n      Ruas, Terry  and\n      Kirstein, Frederic  and\n      Gipp, Bela\",\n    booktitle = \"Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing\",\n    month = dec,\n    year = \"2022\",\n    address = \"Abu Dhabi, United Arab Emirates\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2022.emnlp-main.62\",\n    pages = \"952--963\",\n    abstract = \"The recent success of large language models for text generation poses a severe threat to academic integrity, as plagiarists can generate realistic paraphrases indistinguishable from original work.However, the role of large autoregressive models in generating machine-paraphrased plagiarism and their detection is still incipient in the literature.This work explores T5 and GPT3 for machine-paraphrase generation on scientific articles from arXiv, student theses, and Wikipedia.We evaluate the detection performance of six automated solutions and one commercial plagiarism detection software and perform a human study with 105 participants regarding their detection performance and the quality of generated examples.Our results suggest that large language models can rewrite text humans have difficulty identifying as machine-paraphrased (53{\\%} mean acc.).Human experts rate the quality of paraphrases generated by GPT-3 as high as original texts (clarity 4.0/5, fluency 4.2/5, coherence 3.8/5).The best-performing detection model (GPT-3) achieves 66{\\%} F1-score in detecting paraphrases.We make our code, data, and findings publicly available to facilitate the development of detection solutions.\",\n}\n```\n## License\nThis repository is licensed under the Apache License 2.0 - see the [LICENSE](LICENSE) file for details.\nUse the code for any of your research projects, but be nice and give credit where credit is due.\nAny illegal use for plagiarism or other purposes is prohibited.\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fjpwahle%2Femnlp22-transforming","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fjpwahle%2Femnlp22-transforming","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fjpwahle%2Femnlp22-transforming/lists"}