emotion

A first neuron-level interpretability study of how large audio-language models encode multilingual emotion, introducing Consistency-Regularized Fusion to identify Multilingual Emotion Neurons across 12 languages and showing cross-lingual transfer benefits.

This paper proposes an expanded formulation of valence and arousal in an active inference model of human driving, conditioning affective estimates on predicted future outcomes and evaluating it in interactive driving scenarios.

A new fMRI study in iScience finds that dogs process happy human facial expressions in brain regions linked to higher cognition and reward, and can distinguish between negative expressions like anger, sadness, and fear.

This paper studies how prompt-conditioned emotions affect LLM-based price negotiation, finding that emotions strongly shape outcomes—angry buyers almost never reach agreement while happy buyers agree more often but get worse prices—raising concerns for emotion-conditioned agents in commerce.

This paper proposes using Natural Semantic Metalanguage (NSM) primes as a more basic explanation for emotions in LLMs, showing they are recoverable, causally controllable, and faithful, outperforming appraisal-based directions.

This paper investigates whether induced emotions can bias the sequential decision-making of LLMs using the Iowa Gambling Task as a testbed. The authors find that while emotional induction does not significantly affect average decision dynamics, anger can reduce penalty sensitivity and early-stage exploration.

This paper presents an extended evacuation framework integrating cognitive, emotional, social, and personality mechanisms for agent-based simulations of human behavior under uncertainty. It models dynamic event awareness, memory, fear, and OCEAN-based personality, demonstrating impacts on evacuation efficiency and realistic crowd phenomena.

Introduces MMEE, a multilingual multi-emotion emphasis corpus of 10,000 utterances across 7 languages and 34 emotions, and benchmarks emphasis detection models under various transfer settings, finding that multilingual training improves robustness while monolingual models show limited zero-shot transfer.

EMORSION presents a proof-of-concept study on how film audio parameters (frequency, dynamics, directionality) affect audience emotion and immersion in a cinema setting, finding measurable differences across mixes.

This paper discovers a shared valence axis (V-axis) across modern LLMs and human EEG signals, showing that a single direction from LLM internal representations aligns with neural responses to emotional stimuli. It also identifies the saturation regularity, explaining why LLM-derived supervision fails to improve EEG decoding and how leveraging residual diversity boosts performance.