English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

With ChatGPT under the spotlight, utilizing large language models (LLMs) foracademic writing has drawn a significant amount of discussions and concerns inthe community. While substantial research efforts have been stimulated fordetecting LLM-Generated Content (LLM-content), most of the attempts are stillin the early stage of exploration. In this paper, we present a holisticinvestigation of detecting LLM-generate academic writing, by providing adataset, evidence, and algorithms, in order to inspire more community effort toaddress the concern of LLM academic misuse. We first present GPABenchmark, abenchmarking dataset of 600,000 samples of human-written, GPT-written,GPT-completed, and GPT-polished abstracts of research papers in CS, physics,and humanities and social sciences (HSS). We show that existing open-source andcommercial GPT detectors provide unsatisfactory performance on GPABenchmark,especially for GPT-polished text. Moreover, through a user study of 150+participants, we show that it is highly challenging for human users, includingexperienced faculty members and researchers, to identify GPT-generatedabstracts. We then present CheckGPT, a novel LLM-content detector consisting ofa general representation module and an attentive-BiLSTM classification module,which is accurate, transferable, and interpretable. Experimental results showthat CheckGPT achieves an average classification accuracy of 98% to 99% for thetask-specific discipline-specific detectors and the unified detectors. CheckGPTis also highly transferable that, without tuning, it achieves ~90% accuracy innew domains, such as news articles, while a model tuned with approximately2,000 samples in the target domain achieves ~98% accuracy. Finally, wedemonstrate the explainability insights obtained from CheckGPT to reveal thekey behaviors of how LLM generates texts.

Check Me If You Can: Detecting ChatGPT-Generated Academic Writing using CheckGPT