Search papers, labs, and topics across Lattice.
This paper investigates emotion not as a prediction target, but as a latent factor influencing how LLMs process text. The authors analyze how emotional tone affects attention geometry in transformers, finding correlations between attention metrics (locality, center-of-mass distance, entropy) and question-answering performance. They introduce AURA-QA, a new emotionally-balanced QA dataset, and propose an emotional regularization framework that improves reading comprehension across various benchmarks by constraining emotion-conditioned representational drift.
LLMs' attention patterns subtly shift with emotional tone, and explicitly accounting for these shifts during training improves reading comprehension even on neutral datasets.
Large language models are routinely deployed on text that varies widely in emotional tone, yet their reasoning behavior is typically evaluated without accounting for emotion as a source of representational variation. Prior work has largely treated emotion as a prediction target, for example in sentiment analysis or emotion classification. In contrast, we study emotion as a latent factor that shapes how models attend to and reason over text. We analyze how emotional tone systematically alters attention geometry in transformer models, showing that metrics such as locality, center-of-mass distance, and entropy vary across emotions and correlate with downstream question-answering performance. To facilitate controlled study of these effects, we introduce Affect-Uniform ReAding QA (AURA-QA), a question-answering dataset with emotionally balanced, human-authored context passages. Finally, an emotional regularization framework is proposed that constrains emotion-conditioned representational drift during training. Experiments across multiple QA benchmarks demonstrate that this approach improves reading comprehension in both emotionally-varying and non-emotionally varying datasets, yielding consistent gains under distribution shift and in-domain improvements on several benchmarks.
