Evidence item

The ICLR study describes pretrained language models trained on broad text data and adapted using psychological-experiment data to model human behavior, supporting the limited empirical premise that such systems are embedded in and can model human behavioral data. DIRECT_TEXT

Authenticity UNVERIFIED · extraction UNVERIFIED · independence group none (unreviewed) · created seq 9939

Source: Turning Large Language Models into Cognitive Models · locator {} · excerpt hash sha256:66a58a36a462c6937f74b679b9e61d73af512642665c2f3dfaacd0e829e5efae

Large language models are neural networks trained on vast amounts of data to predict the next token for a given text sequence. ... after finetuning them on data from psychological experiments ... these models offer accurate representations of human behavior ... Taken together, these results suggest that large, pre-trained models can be adapted to become models of human cognition.