petersuber on Nostr: Nice #AI test. Researchers put real #neuroscience abstracts up against copies that ...

Nice #AI test. Researchers put real #neuroscience abstracts up against copies that described the same question and methods but (deliberately) different results. Then asked 15 #LLMs and 171 human neuroscientists to tell which was which. "All of the LLMs outperformed the neuroscientists, with the LLMs averaging 81% accuracy and the humans averaging 63% accuracy."

* summary
https://medicalxpress.com/news/2024-11-ai-neuroscience-results-human-experts.html
* primary source
https://www.nature.com/articles/s41562-024-02046-9