Sparse Autoencoders for Interpretable Emotion Control in Text-to-Speech
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866916072088666112 |
|---|---|
| author | Du, Hongfei Shi, Jiacheng Lu, Sidi Zhou, Gang Gao, Ye |
| author_facet | Du, Hongfei Shi, Jiacheng Lu, Sidi Zhou, Gang Gao, Ye |
| contents | Integrating large language models (LLMs) into text-to-speech (TTS) systems has improved speech expressiveness, yet interpretable emotional control remains challenging. Existing approaches primarily rely on external conditioning or global activation steering, offering limited insight into the internal representations underlying emotional control. In this work, we analyze emotion-related variation in the semantic hidden states of LLM-based TTS models using sparse autoencoders (SAEs) to identify sparse latent features. Our analysis shows that emotional variation is distributed across multiple sparse latent features, while intervening on a small subset enables interpretable emotion control. Building on this observation, we introduce a feature-level intervention framework for bidirectional emotion induction and suppression without modifying backbone parameters. We further show that distinct latent features are associated with specific acoustic attributes (e.g., pitch), suggesting that emotional expression arises from coordinated latent contributions rather than a single global shift. Empirically, steering these sparse latent features achieves comparable or superior emotion induction and suppression performance relative to global steering and existing TTS baselines. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2606_01479 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Sparse Autoencoders for Interpretable Emotion Control in Text-to-Speech Du, Hongfei Shi, Jiacheng Lu, Sidi Zhou, Gang Gao, Ye Computation and Language Integrating large language models (LLMs) into text-to-speech (TTS) systems has improved speech expressiveness, yet interpretable emotional control remains challenging. Existing approaches primarily rely on external conditioning or global activation steering, offering limited insight into the internal representations underlying emotional control. In this work, we analyze emotion-related variation in the semantic hidden states of LLM-based TTS models using sparse autoencoders (SAEs) to identify sparse latent features. Our analysis shows that emotional variation is distributed across multiple sparse latent features, while intervening on a small subset enables interpretable emotion control. Building on this observation, we introduce a feature-level intervention framework for bidirectional emotion induction and suppression without modifying backbone parameters. We further show that distinct latent features are associated with specific acoustic attributes (e.g., pitch), suggesting that emotional expression arises from coordinated latent contributions rather than a single global shift. Empirically, steering these sparse latent features achieves comparable or superior emotion induction and suppression performance relative to global steering and existing TTS baselines. |
| title | Sparse Autoencoders for Interpretable Emotion Control in Text-to-Speech |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2606.01479 |