Language Models Don't Always Say What They Think: Unfaithful Explanations in Chain-of-Thought Prompting
Samuel Bowman, Julian Michael, Ethan Perez, Miles Turpin
Topics & Concepts
PsychologyCognitive psychologyLinguisticsSocial psychologyComputer scienceAction (physics)Perspective (graphical)Language modelIdentification (biology)Cognitive sciencePerceptionEpistemologyCommunicationNatural languageNarrativeCausality (physics)Embodied and Extended CognitionTopic ModelingEpistemology, Ethics, and Metaphysics