Showing 1–1 of 1 results for author: Tak, A N

Is GPT a Computational Model of Emotion? Detailed Analysis

Authors: Ala N. Tak, Jonathan Gratch

Abstract: This paper investigates the emotional reasoning abilities of the GPT family of large language models via a component perspective. The paper first examines how the model reasons about autobiographical memories. Second, it systematically varies aspects of situations to impact emotion intensity and co** tendencies. Even without the use of prompt engineering, it is shown that GPT's predictions align… ▽ More This paper investigates the emotional reasoning abilities of the GPT family of large language models via a component perspective. The paper first examines how the model reasons about autobiographical memories. Second, it systematically varies aspects of situations to impact emotion intensity and co** tendencies. Even without the use of prompt engineering, it is shown that GPT's predictions align significantly with human-provided appraisals and emotional labels. However, GPT faces difficulties predicting emotion intensity and co** responses. GPT-4 showed the highest performance in the initial study but fell short in the second, despite providing superior results after minor prompt engineering. This assessment brings up questions on how to effectively employ the strong points and address the weak areas of these models, particularly concerning response variability. These studies underscore the merits of evaluating models from a componential perspective. △ Less

Submitted 25 July, 2023; originally announced July 2023.