Really interesting new paper exploring how GTP-generated sentences can drive brain region specific responses in humans reading those sentences. Could be a first glimpse into better understanding how LLMs-AI models "interact" with human language as perceived by the brain (in simplified terms). It's with restricted access on Nature but the preprint can be downloaded under https://www.biorxiv.org/content/10.1101/2023.04.16.537080v1

#science #neuroscience #neurology #AI #LLM