Explore how large language models represent rhetorical questions using linear probing across social media datasets, revealing distinct linguistic cues.
Discover how reinforcement learning enhances sentiment analysis by aligning AI reasoning with human rationale for better accuracy and interpretability.