Language-specific representation of emotion-concept knowledge causally supports emotion inference
Ming Li,
Yusheng Su,
Hsiu-Yuan Huang,
Jiali Cheng,
Xin Hu,
Xinmiao Zhang,
Huadong Wang,
Yujia Qin,
Xiaozhi Wang,
Kristen A. Lindquist,
Zhiyuan Liu,
Dan Zhang
Affiliations
Ming Li
Department of Psychological and Cognitive Sciences, Tsinghua University, Beijing, China; Tsinghua Laboratory of Brain and Intelligence, Tsinghua University, Beijing, China
Yusheng Su
Department of Computer Science and Technology, Tsinghua University, Beijing, China
Hsiu-Yuan Huang
School of Computer and Communication Engineering, University of Science and Technology Beijing, Beijing, China
Jiali Cheng
Miner School of Computer and Information Sciences, University of Massachusetts Lowell, Lowell, MA 01854, USA
Xin Hu
Department of Psychiatry, University of Pittsburgh, Pittsburgh, PA 15260, USA
Xinmiao Zhang
Department of Psychological and Cognitive Sciences, Tsinghua University, Beijing, China; Tsinghua Laboratory of Brain and Intelligence, Tsinghua University, Beijing, China
Huadong Wang
Department of Computer Science and Technology, Tsinghua University, Beijing, China
Yujia Qin
Department of Computer Science and Technology, Tsinghua University, Beijing, China
Xiaozhi Wang
Department of Computer Science and Technology, Tsinghua University, Beijing, China
Kristen A. Lindquist
Department of Psychology and Neuroscience, University of North Carolina, Chapel Hill, NC 27599, USA
Zhiyuan Liu
Department of Computer Science and Technology, Tsinghua University, Beijing, China; Corresponding author
Dan Zhang
Department of Psychological and Cognitive Sciences, Tsinghua University, Beijing, China; Tsinghua Laboratory of Brain and Intelligence, Tsinghua University, Beijing, China; Corresponding author
Summary: Humans no doubt use language to communicate about their emotional experiences, but does language in turn help humans understand emotions, or is language just a vehicle of communication? This study used a form of artificial intelligence (AI) known as large language models (LLMs) to assess whether language-based representations of emotion causally contribute to the AI’s ability to generate inferences about the emotional meaning of novel situations. Fourteen attributes of human emotion concept representation were found to be represented by the LLM’s distinct artificial neuron populations. By manipulating these attribute-related neurons, we in turn demonstrated the role of emotion concept knowledge in generative emotion inference. The attribute-specific performance deterioration was related to the importance of different attributes in human mental space. Our findings provide a proof-in-concept that even an LLM can learn about emotions in the absence of sensory-motor representations and highlight the contribution of language-derived emotion-concept knowledge for emotion inference.
