Eine neue Studie analysierte über 1 Million Forschungsarbeiten und ergab eine wachsende Verwendung von großsprachigen Modellen, insbesondere in Informatik, wo bis zu 22% Anzeichen einer LLM -Bearbeitung zeigen.
https://www.nature.com/articles/s41562-025-02273-8
3 Kommentare
You can find a lot of papers which use LLM generated text just simply googling `site:arxiv.org do not highlight any negatives`. This is really concerning.
I dont know man, i had do the excercise of putting my own writing on analizers for iA and have gotten even “80% of probability of use” when i never did it…
Anyone got the article? Can’t access this via my institution 🙁