{"type":"conference","conference":{"start_date":"2024-11-12","name":"2024 Conference on Empirical Methods in Natural Language Processing (EMNLP 2024)","location":"Miami, FL, USA","end_date":"2024-11-16"},"date_created":"2025-09-11T07:11:34Z","language":[{"iso":"eng"}],"doi":"10.18653/v1/2024.emnlp-main.564","place":"Miami, FL, USA","department":[{"_id":"660"}],"author":[{"last_name":"Xu","full_name":"Xu, Yang","first_name":"Yang"},{"first_name":"Yu","full_name":"Wang, Yu","last_name":"Wang"},{"last_name":"An","full_name":"An, Hao","first_name":"Hao"},{"first_name":"Zhichen","last_name":"Liu","full_name":"Liu, Zhichen"},{"first_name":"Yongyuan","full_name":"Li, Yongyuan","last_name":"Li"}],"status":"public","oa":"1","publication":"Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing","main_file_link":[{"open_access":"1","url":"https://doi.org/10.18653/v1/2024.emnlp-main.564"}],"year":"2024","_id":"61177","citation":{"short":"Y. Xu, Y. Wang, H. An, Z. Liu, Y. Li, in: Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, ACL, Miami, FL, USA, 2024, pp. 10108–10121.","chicago":"Xu, Yang, Yu Wang, Hao An, Zhichen Liu, and Yongyuan Li. “Detecting Subtle Differences between Human and Model Languages Using Spectrum of Relative Likelihood.” In <i>Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing</i>, 10108–10121. Miami, FL, USA: ACL, 2024. <a href=\"https://doi.org/10.18653/v1/2024.emnlp-main.564\">https://doi.org/10.18653/v1/2024.emnlp-main.564</a>.","apa":"Xu, Y., Wang, Y., An, H., Liu, Z., &#38; Li, Y. (2024). Detecting subtle differences between human and model languages using spectrum of relative likelihood. <i>Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing</i>, 10108–10121. <a href=\"https://doi.org/10.18653/v1/2024.emnlp-main.564\">https://doi.org/10.18653/v1/2024.emnlp-main.564</a>","ama":"Xu Y, Wang Y, An H, Liu Z, Li Y. Detecting subtle differences between human and model languages using spectrum of relative likelihood. In: <i>Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing</i>. ACL; 2024:10108–10121. doi:<a href=\"https://doi.org/10.18653/v1/2024.emnlp-main.564\">10.18653/v1/2024.emnlp-main.564</a>","mla":"Xu, Yang, et al. “Detecting Subtle Differences between Human and Model Languages Using Spectrum of Relative Likelihood.” <i>Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing</i>, ACL, 2024, pp. 10108–10121, doi:<a href=\"https://doi.org/10.18653/v1/2024.emnlp-main.564\">10.18653/v1/2024.emnlp-main.564</a>.","ieee":"Y. Xu, Y. Wang, H. An, Z. Liu, and Y. Li, “Detecting subtle differences between human and model languages using spectrum of relative likelihood,” in <i>Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing</i>, Miami, FL, USA, 2024, pp. 10108–10121, doi: <a href=\"https://doi.org/10.18653/v1/2024.emnlp-main.564\">10.18653/v1/2024.emnlp-main.564</a>.","bibtex":"@inproceedings{Xu_Wang_An_Liu_Li_2024, place={Miami, FL, USA}, title={Detecting subtle differences between human and model languages using spectrum of relative likelihood}, DOI={<a href=\"https://doi.org/10.18653/v1/2024.emnlp-main.564\">10.18653/v1/2024.emnlp-main.564</a>}, booktitle={Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing}, publisher={ACL}, author={Xu, Yang and Wang, Yu and An, Hao and Liu, Zhichen and Li, Yongyuan}, year={2024}, pages={10108–10121} }"},"project":[{"name":"TRR 318; TP A02: Verstehensprozess einer Erklärung beobachten und auswerten","_id":"112"}],"date_updated":"2025-09-11T10:15:56Z","abstract":[{"text":"Human and model-generated texts can be distinguished by examining the magnitude of likelihood in language. However, it is becoming increasingly difficult as language model's capabilities of generating human-like texts keep evolving. This study provides a new perspective by using the relative likelihood values instead of absolute ones, and extracting useful features from the spectrum-view of likelihood for the human-model text detection task. We propose a detection procedure with two classification methods, supervised and heuristic-based, respectively, which results in competitive performances with previous zero-shot detection methods and a new state-of-the-art on short-text detection. Our method can also reveal subtle differences between human and model languages, which find theoretical roots in psycholinguistics studies.","lang":"eng"}],"quality_controlled":"1","extern":"1","publication_status":"published","publisher":"ACL","user_id":"76456","page":"10108–10121","title":"Detecting subtle differences between human and model languages using spectrum of relative likelihood"}