# Multilingual Emotion Neurons in Large Audio-Language Models

arXiv: https://arxiv.org/abs/2608.08772 · HTML: https://mlsanity.com/paper/2608.08772/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 25 of its 56 references that are in the map.
- First posted on arXiv 9 Aug 2026.

## Builds on (most cited first, 25 of 25)

- [Sparse Autoencoders Find Highly Interpretable Features in Language Models](https://mlsanity.com/paper/2309.08600/) (2023)
- [Network Dissection: Quantifying Interpretability of Deep Visual Representations](https://mlsanity.com/paper/1704.05796/) (2017)
- [Six Challenges for Neural Machine Translation](https://mlsanity.com/paper/1706.03872/) (2017)
- [MMAU: A Massive Multi-Task Audio Understanding and Reasoning Benchmark](https://mlsanity.com/paper/2410.19168/) (2024)
- [Style Tokens: Unsupervised Style Modeling, Control and Transfer in End-to-End Speech Synthesis](https://mlsanity.com/paper/1803.09017/) (2018)
- [Understanding the Role of Individual Units in a Deep Neural Network](https://mlsanity.com/paper/2009.05041/) (2020)
- [Towards End-to-End Prosody Transfer for Expressive Speech Synthesis with Tacotron](https://mlsanity.com/paper/1803.09047/) (2018)
- [What Is One Grain of Sand in the Desert? Analyzing Individual Neurons in Deep NLP Models](https://mlsanity.com/paper/1812.09355/) (2018)
- [Emotion-LLaMA: Multimodal Emotion Recognition and Reasoning with Instruction Tuning](https://mlsanity.com/paper/2406.11161/) (2024)
- [Identifying and Controlling Important Neurons in Neural Machine Translation](https://mlsanity.com/paper/1811.01157/) (2018)
- [Universal Neurons in GPT2 Language Models](https://mlsanity.com/paper/2401.12181/) (2024)
- [Dynamic-SUPERB: Towards A Dynamic, Collaborative, and Comprehensive Instruction-Tuning Benchmark for Speech](https://mlsanity.com/paper/2309.09510/) (2023)
- [Neurons in Large Language Models: Dead, N-gram, Positional](https://mlsanity.com/paper/2309.04827/) (2023)
- [EmoBox: Multilingual Multi-corpus Speech Emotion Recognition Toolkit and Benchmark](https://mlsanity.com/paper/2406.07162/) (2024)
- [Expressive TTS Training with Frame and Style Reconstruction Loss](https://mlsanity.com/paper/2008.01490/) (2020)
- [Fine-grained Emotion Strength Transfer, Control and Prediction for Emotional Speech Synthesis](https://mlsanity.com/paper/2011.08477/) (2020)
- [EmoSteer-TTS: Fine-Grained and Training-Free Emotion-Controllable Text-to-Speech via Activation Steering](https://mlsanity.com/paper/2508.03543/) (2025)
- [ShEMO -- A Large-Scale Validated Database for Persian Speech Emotion Detection](https://mlsanity.com/paper/1906.01155/) (2019)
- [End-to-End Emotional Speech Synthesis Using Style Tokens and Semi-Supervised Training](https://mlsanity.com/paper/1906.10859/) (2019)
- [Multimodal Large Language Models Meet Multimodal Emotion Recognition and Reasoning: A Survey](https://mlsanity.com/paper/2509.24322/) (2025)
- [Do LLMs "Feel"? Emotion Circuits Discovery and Control](https://mlsanity.com/paper/2510.11328/) (2025)
- [Cross-lingual and Multilingual Speech Emotion Recognition on English and French](https://mlsanity.com/paper/1803.00357/) (2018)
- [A New Amharic Speech Emotion Dataset and Classification Benchmark](https://mlsanity.com/paper/2201.02710/) (2022)
- [nEMO: Dataset of Emotional Speech in Polish](https://mlsanity.com/paper/2404.06292/) (2024)
- [Cross-Lingual Speech Emotion Recognition: Humans vs. Self-Supervised Models](https://mlsanity.com/paper/2409.16920/) (2024)

## Built on by (most cited first, 1 of 1)

- [Emotion Across Speech and Faces: Shared Affective Mechanisms in Multimodal Foundation Models](https://mlsanity.com/paper/2608.17102/) (2026)

## Abstract (arXiv)

Emotion is central to human communication, and its expression varies across languages. Large audio-language models (LALMs) achieve strong performance on multilingual speech tasks, yet it remains unclear whether they encode emotion through language-specific correlations or language-agnostic representations. We present the first neuron-level interpretability study of this question. We define Multilingual Emotion Neurons (MLENs) as functional units exhibiting stable emotional selectivity and aligned causal effects across languages, and introduce Consistency-Regularized Fusion (CR-Fusion) to identify them. Across four modern LALMs and 12 typologically diverse languages, emotion-sensitive neurons identified independently per language show minimal overlap, and additional monolingual identification data saturates quickly without isolating more transferable units, motivating identification from pooled cross-lingual evidence. Causal interventions demonstrate that MLENs identified by CR-Fusion provide more precise and transferable affective control than monolingual neuron sets in both zero-shot and low-resource settings. Leave-one-out ablations further reveal asymmetric transfer: individual identification languages, including low-resource ones, contribute non-redundant evidence, while several low-resource languages benefit most from the resulting cross-lingual transfer. Together, our findings provide the first causal, neuron-level account of how LALMs encode emotion across languages, and establish multilingual neuron identification as an effective mechanism for understanding cross-lingual affective behavior.