Behavioral Emotion Analysis Model for Large Language Models

In human-computer interaction, recognizing and responding to a user's emotional state is crucial for effective communication and successful task completion. For instance, a caregiving AI agent capable of detecting pain or depression in a patient could offer tailored empathetic support and appropriate medical interventions while adhering to ethical guidelines and safeguarding patient well-being. This paper examines cognitive research on human emotions and proposes the Behavioral Emotion Analysis Model (BEAM), a novel emotion spectrum framework that incorporates both basic emotions and their linguistic antonyms. BEAM provides a comprehensive way to understand and represent emotional states in language and is designed to be integrated with Large Language Models (LLMs). By leveraging BEAM, LLMs can adapt their linguistic behaviors and expressions based on the detected emotional state of the user, ensuring responses are both empathetic and ethically aligned.