Emotion understanding in discourse requires reasoning beyond surface sentiment because speakers often convey affect through indirect, implicit, polite, ironic, or deliberately mismatched expressions. Existing emotion benchmarks mainly annotate surface polarity or final emotion categories, while lacking a structured account of how explicit expression, implicit affect, pragmatic intent, and fine grained emotion interact. This limitation makes current evaluations insensitive to cases where affective meaning is concealed, weakened, inverted, or pragmatically reshaped, thereby obscuring model failures in deeper emotion understanding. To address this gap, we introduce CUE Bench, a Chinese Unsaid Emotion benchmark that centers on Affective Stance and covers diverse communicative scenarios. CUE Bench constructs nine human interpretable affective stances from explicit implicit polarity interaction and further provides intent and fine grained emotion annotations for structured affective inference. Experiments show that incorporating Affective Stance improves fine grained emotion recognition by 3.1 percentage points and pragmatic intent detection by 8.1 percentage points over strong baselines.
Empathetic response generation requires models to decide not only what to say, but also how to respond to the previous speaker's affective situation, so multi-annotator emoji distributions are used as weak affective--attitudinal evidence to induce a latent control space that operationally approximates listener stance.
An LLM-as-Judge framework is introduced that evaluates each emotion independently according to its plausibility in the conversational context rather than enforcing a single-label decision, suggesting that standard single-label evaluation is therefore insufficient.
Amir Ben Khalifa, Fanny Bezancon, B. Abdulrazak et al.· 0 citations
This work constructs the MaskDialog dataset carefully curated from television drama and large language models, and proposes two LLM-based baseline approaches, i.e., One-shot Self-consistent Inference and Cascaded Multi-step Inference, and conducts comprehensive analyses on dialogue construction strategies and inference...
Zhi-Qiang Gao, Jing Han, Zhuo-Chu Wang et al.· Proceedings of the Thirty-Fi...· 0 citations
This work introduces CHIARO, a 1,000 human-annotated sentence benchmark for contrastive emotion inference grounded in appraisal theory, and benchmarked seven frontier LLMs and four off-the-shelf emotion classifiers.
Divyesh Bommana, Mohammad Saim, Tianyu Jiang· 0 citations
While multimodal large language models (MLLMs) have demonstrated exceptional capabilities in objective understanding tasks, their performance in affective reasoning still falls significantly short of human standards. We attribute it to a central capability gap: MLLMs are difficult to reliably distinguish semantically p...
C. Ye, Wei-Dong Chen, Zhao-Bo Qi et al.· 0 citations
Pure expressive adjectives (EAs) (e.g., damn) convey a negative speaker attitude toward a discourse referent and
display interpretational flexibility, often targeting elements beyond their syntactic sister. The speech-act hypothesis treats EAs
as quasi-independent, non-at-issue speech acts whose target is pragmatic...
Kalle Glauch· Pragmatics & Cognition· 0 citations
