Публикация

We often interpret ambiguous feedback in ways that confirm what we already believe. A laugh can feel supportive or mocking. A neutral email can feel encouraging or disappointing. A new paper studies how this happens during reward learning.