Elamur Research⌕
☰
home / papers / 2026 / SearchLLM: Detecting LLM Paraphrased Text by Measuring the Similarity with Regeneration of the Candidate Source via Search Engine

SearchLLM: Detecting LLM Paraphrased Text by Measuring the Similarity with Regeneration of the Candidate Source via Search Engine

Summary

When people let an LLM polish their drafts, the meaning can drift, and the result is often hard to tell apart from human writing. Existing watermarking, supervised and zero-shot detectors struggle most when the paraphrase stays very close to the original. The authors observe that the same LLM tends to produce nearly identical paraphrases of a given text across repeated runs. SearchLLM builds on this. It uses a search engine (Google Search API, or the Wikipedia API) to find a likely source for the input text.

  • If the input closely matches that source above a threshold α, it is labelled human-written.
  • Otherwise, the LLM rewrites the candidate source, and high similarity between this rewrite and the input points to LLM paraphrasing.
  • If neither check is conclusive, the decision passes to an existing detector.

Because of this design, SearchLLM works as a proxy layer on top of any detector. Experiments on MAGE and XSum show consistent gains for recent detectors, especially on close-mimic paraphrases, along with better robustness to paraphrasing attacks such as DIPPER.

Key points

  • Motivating study: 1,000 human and 1,000 LLM-paraphrased samples were compared with sources retrieved from the web. Human texts were usually more similar to their sources. Regeneration with GPT-4o-mini moved similarity in opposite directions for the two classes.
  • Evaluation data:
    • 300 human and GPT-3.5-turbo texts sampled from MAGE.
    • 150 XSum articles paraphrased with GPT-4o-mini and GPT-4o, plus revised and polished variants.
    • Additional results on the Dugan et al. (2024) dataset and on low-resource languages from M4.
  • Main finding: Baseline detectors, including the MAGE-trained Longformer, do well on MAGE but drop sharply on close-mimic XSum paraphrases. SearchLLM recovers much of this loss and helps further on short texts.
  • Cost: Roughly 50% of texts yield a usable candidate source. The total cost is under $0.00515 per sample, which is cheaper than commercial LLMs with built-in web search.
  • Limitations and code: The method only works with public, indexed sources and struggles with freely generated or cross-lingual text. It is also exposed to adaptive attacks that plant manipulated web content. Code is available at github.com/quocnsh/SearchLLM.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

Fields: Machine translation · Natural language processing

Cite

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.eacl-long.79

@inproceedings{nguyenson2026searchllm,
  title = {{SearchLLM: Detecting LLM Paraphrased Text by Measuring the Similarity with Regeneration of the Candidate Source via Search Engine}},
  author = {Nguyen-Son, Hoang-Quoc and Dao, Minh-Son and Zettsu, Koji},
  booktitle = {Proceedings of the 19th Conference of the European Chapter of the Association for Computational Linguistics (Volume 1: Long Papers)},
  year = {2026},
  doi = {10.18653/v1/2026.eacl-long.79},
  url = {https://doi.org/10.18653/v1/2026.eacl-long.79}
}

References

39 references, 5 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Guangsheng Bao, Yanbin Zhao, Juncai He, and Yue Zhang. 2025. Glimpse: Enabling white-box methods to use proprietary models for zero-shot llm-generated text detection. In Proceedings of the 13rd International Conference on Learning Representations (ICLR). ↗
  2. Guangsheng Bao, Yanbin Zhao, Zhiyang Teng, Linyi Yang, and Yue Zhang. 2024. Fast-detectgpt: Efficient zero-shot detection of machine-generated text via conditional probability curvature. In Proceedings of the 12th International Conference on Learning Representations (ICLR). ↗ our card
  3. Amrita Bhattacharjee and Huan Liu. 2024. Fighting fire with fire: can chatgpt detect ai-generated text? In Proceedings of the 30th ACM Conference on Knowledge Discovery and Data Mining Explorations Newsletter (SIGKDD), volume 25, pages 14–21. ↗
  4. Yapei Chang, Kalpesh Krishna, Amir Houmansadr, John Wieting, and Mohit Iyyer. 2024. Postmark: A robust blackbox watermark for large language models. In Proceedings of the Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 8969–8987. find
  5. Liam Dugan, Alyssa Hwang, Filip Trhlík, Andrew Zhu, Josh Magnus Ludan, Hainiu Xu, Daphne Ippolito, and Chris Callison-Burch. 2024. Raid: A shared benchmark for robust evaluation of machine-generated text detectors. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (ACL), pages 12463–12492. detection by checking memorization of preceding tokens. In Proceedings of the 37th Annual Conference on Neural Information Processing Systems (NeurIPS), volume 37, pages 104065–104090. ↗
  6. Xun Guo, Yongxin He, Shan Zhang, Ting Zhang, Wanquan Feng, Haibin Huang, and Chongyang Ma. 2024b. Detective: Detecting ai-generated text via multi-level contrastive learning. In Proceedings of the 38th Annual Conference on Neural Information Processing Systems (NeurIPS), volume 37, pages 88320–88347. find
  7. Abhimanyu Hans, Avi Schwarzschild, Valeriia Cherepanova, Hamid Kazemi, Aniruddha Saha, Micah Goldblum, Jonas Geiping, and Tom Goldstein. 2024. Spotting llms with binoculars: zero-shot detection of machine-generated text. In Proceedings of the 41st International Conference on Machine Learning (ICML), pages 17519–17537. ↗
  8. Xiaomeng Hu, Pin-Yu Chen, and Tsung-Yi Ho. 2023. Radar: Robust ai-text detection via adversarial learning. In Proceedings of the 37th Annual Conference on Neural Information Processing Systems (NeurIPS), volume 36, pages 15077–15095. find
  9. Mingjia Huo, Sai Ashish Somayajula, Youwei Liang, Ruisi Zhang, Farinaz Koushanfar, and Pengtao Xie. 2024. Token-specific watermarking with enhanced detectability and semantic coherence for large language models. In Proceedings of the International Conference on Machine Learning (ICML), pages 20746–20767. ↗
  10. Zae Myung Kim, Kwang Hee Lee, Preston Zhu, Vipul Raheja, and Dongyeop Kang. 2024. Threads of subtlety: Detecting machine-generated texts through discourse motifs. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (ACL), pages 5449–5474. find
  11. John Kirchenbauer, Jonas Geiping, Yuxin Wen, Jonathan Katz, Ian Miers, and Tom Goldstein. 2023. A watermark for large language models. In Proceedings of the International Conference on Machine Learning (ICML), pages 17061–17084. ↗ our card
  12. Ryuto Koike, Masahiro Kaneko, and Naoaki Okazaki. 2024. Outfox: Llm-generated essay detection through in-context learning with adversarially generated examples. In Proceedings of the Conference on Artificial Intelligence (AAAI), volume 38, pages 21258–21266. find
  13. Thibaud Gloaguen, Nikola Jovanović, Robin Staab, and Martin Vechev. 2025. Black-box detection of language model watermarks. In Proceedings of the International Conference on Learning Representations (ICLR). ↗
  14. Kalpesh Krishna, Yixiao Song, Marzena Karpinska, John Wieting, and Mohit Iyyer. 2023. Paraphrasing evades detectors of ai-generated text, but retrieval is an effective defense. In Proceedings of the 37th Annual Conference on Neural Information Processing Systems (NeurIPS), volume 36, pages 27469–27500. find our card
  15. Hanxi Guo, Siyuan Cheng, Xiaolong Jin, Zhuo Zhang, Kaiyuan Zhang, Guanhong Tao, Guangyu Shen, and Xiangyu Zhang. 2024a. Biscope: Ai-generated text Xiang Li, Zhiyi Yin, Hexiang Tan, Shaoling Jing, Du Su, Yi Cheng, Huawei Shen, and Fei Sun. 2025. Prde-tect: Perturbation-robust llm-generated text detection based on syntax tree. In Findings of the Association for Computational Linguistics (NAACL), pages 8290–8301. find
  16. Yafu Li, Qintong Li, Leyang Cui, Wei Bi, Zhilin Wang, Longyue Wang, Linyi Yang, Shuming Shi, and Yue Zhang. 2024. Mage: Machine-generated text detection in the wild. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (ACL), pages 36–53. ↗
  17. Shengchao Liu, Xiaoming Liu, Yichen Wang, Zehua Cheng, Chengzhengxu Li, Zhaohan Zhang, Yu Lan, and Chao Shen. 2024. Does detectgpt fully utilize perturbation? bridging selective perturbation to fine-tuned contrastive learning detector would be better. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (ACL), pages 1874–1889. ↗ our card
  18. Xiaoming Liu, Zhaohan Zhang, Yichen Wang, Hang Pu, Yu Lan, and Chao Shen. 2023. Coco: Coherence-enhanced machine-generated text detection under low resource with contrastive learning. In Proceedings of the Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 16167–16188. find
  19. Yijian Lu, Aiwei Liu, Dianzhi Yu, Jingjing Li, and Irwin King. 2024. An entropy-based text watermarking detection method. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (ACL), pages 11724–11735. find
  20. Shixuan Ma and Quan Wang. 2024. Zero-shot detection of llm-generated text using token cohesiveness. In Proceedings of the Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 17538–17553. ↗
  21. Chengzhi Mao, Carl Vondrick, Hao Wang, and Junfeng Yang. 2024. Raidar: generative ai detection via rewriting. In Proceedings of the 12th International Conference on Learning Representations (ICLR). ↗
  22. Niloofar Mireshghallah, Justus Mattern, Sicun Gao, Reza Shokri, and Taylor Berg-Kirkpatrick. 2024. Smaller language models are better zero-shot machine-generated text detectors. In Proceedings of the 18th Conference of the European Chapter of the Association for Computational Linguistics (EACL), pages 278–293. find
  23. Hoang-Quoc Nguyen-Son, Minh-Son Dao, and Koji Zettsu. 2024. Simllm: Detecting sentences generated by large language models using similarity between the generation and its re-generation. In Proceedings of the Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 22340– 22352. find
  24. Hyeonchu Park, Byungjun Kim, and Bugeun Kim. 2025. Dart: An aigt detector using amr of rephrased text. In Proceedings of the Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics (NAACL), pages 710–721. ↗
  25. Yuhui Shi, Qiang Sheng, Juan Cao, Hao Mi, Beizhe Hu, and Danding Wang. 2024. Ten words only still help: improving black-box ai-generated text detection via proxy-guided efficient re-sampling. In Proceedings of the 33rd International Joint Conference on Artificial Intelligence (IJCAI), pages 494–502. find
  26. Jinyan Su, Terry Zhuo, Di Wang, and Preslav Nakov. 2023. Detectllm: Leveraging log rank information for zero-shot detection of machine-generated text. In Findings of the Association for Computational Linguistics (EMNLP), pages 12395–12412. find
  27. Yuchuan Tian, Hanting Chen, Xutao Wang, Zheyuan Bai, Qinghua Zhang, Ruifeng Li, Chao Xu, and Yunhe Wang. 2024. Multiscale positive-unlabeled detection of ai-generated texts. In Proceedings of the International Conference on Learning Representations (ICLR). ↗
  28. Vivek Verma, Eve Fleisig, Nicholas Tomlin, and Dan Klein. 2024. Ghostbuster: Detecting text ghostwritten by large language models. In Proceedings of the Conference of the North American Chapter of the Association for Computational Linguistics (NAACL), pages 1702–1717. ↗
  29. Yuxia Wang, Jonibek Mansurov, Petar Ivanov, Jinyan Su, Artem Shelmanov, Akim Tsvigun, Chenxi Whitehouse, Osama Mohammed Afzal, Tarek Mahmoud, Toru Sasaki, and 1 others. 2024. M4: Multi-generator, multi-domain, and multi-lingual black-box machine-generated text detection. In Proceedings of the 18th Conference of the European Chapter of the Association for Computational Linguistics (ACL), pages 1369–1407. find
  30. Eric Mitchell, Yoonho Lee, Alexander Khazatsky, Christopher D Manning, and Chelsea Finn. 2023. Detectgpt: Zero-shot machine-generated text detection using probability curvature. In Proceedings of the International Conference on Machine Learning (ICML), pages 24950–24962. PMLR. find our card
  31. Kangxi Wu, Liang Pang, Huawei Shen, Xueqi Cheng, and Tat-Seng Chua. 2023. Llmdet: A third party large language models generated text detection tool. In Findings of the Association for Computational Linguistics (EMNLP), pages 2113–2133. find
  32. Shashi Narayan, Shay B Cohen, and Mirella Lapata. 2018. Don’t give me the details, just the summary! topic-aware convolutional neural networks for extreme summarization. In Proceedings of the Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 1797–1807. ↗
  33. Yang Xu, Yu Wang, Hao An, Zhichen Liu, and Yongyuan Li. 2024. Detecting subtle differences between human and model languages using spectrum of relative likelihood. In Proceedings of the Conference on Empirical Methods in Natural Language Processing (EMNLP). find
  34. Yihuai Xu, Yongwei Wang, Yifei Bi, Huangsen Cao, Zhouhan Lin, Yu Zhao, and Fei Wu. 2025. Training-free llm-generated text detection by mining token probability sequences. In Proceedings of the 13rd International Conference on Learning Representations (ICLR). ↗
  35. Xianjun Yang, Wei Cheng, Yue Wu, Linda Ruth Petzold, William Yang Wang, and Haifeng Chen. 2024. Dna-gpt: Divergent n-gram analysis for training-free detection of gpt-generated text. In Proceedings of the 12th International Conference on Learning Representations (ICLR). ↗
  36. Xiao Yu, Kejiang Chen, Qi Yang, Weiming Zhang, and Nenghai Yu. 2024a. Text fluoroscopy: Detecting llm-generated text through intrinsic features. In Proceedings of the Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 15838–15846. find
  37. Xiao Yu, Yuang Qi, Kejiang Chen, Guoqiang Chen, Xi Yang, Pengyuan Zhu, Xiuwei Shang, Weiming Zhang, and Nenghai Yu. 2024b. Dpic: Decoupling prompt and intrinsic characteristics for llm generated text detection. In Proceedings of the 37th Annual Conference on Neural Information Processing Systems (NeurIPS), volume 37, pages 16194–16212. find
  38. Cong Zeng, Shengkun Tang, Xianjun Yang, Yuanzhou Chen, Yiyou Sun, Zhiqiang Xu, Yao Li, Haifeng Chen, Wei Cheng, and Dongkuan DK Xu. 2024. Dald: Improving logits-based detector without logits from black-box llms. In Proceedings of the 38th Annual Conference on Neural Information Processing Systems (NeurIPS), volume 37, pages 54947–54973. find
  39. Biru Zhu, Lifan Yuan, Ganqu Cui, Yangyi Chen, Chong Fu, Bingxiang He, Yangdong Deng, Zhiyuan Liu, Maosong Sun, and Ming Gu. 2023. Beat llms at their own game: Zero-shot llm-generated text detection via querying chatgpt. In Proceedings of the Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 7470–7483. find

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/son-2026-searchllm-detecting-llm-paraphrased-text/