World Model for Sarcasm Detection

This paper addresses the challenging problem of sarcasm understanding in NLP. It proposes a novel approach, WM-SAR, that leverages LLMs and decomposes the reasoning process into specialized agents. The key contribution is the explicit modeling of cognitive factors like literal meaning, context, and intention, leading to improved performance and interpretability compared to black-box methods. The use of a deterministic inconsistency score and a lightweight Logistic Regression model for final prediction is also noteworthy.

• •Proposes WM-SAR, a novel approach to sarcasm detection.
• •Decomposes sarcasm understanding into specialized LLM-based agents.
• •Explicitly models cognitive factors for improved interpretability.
• •Achieves state-of-the-art performance on sarcasm detection benchmarks.