Fetching the paper…

Cognitive models can reveal interpretable value trade-offs in language models · Around