Guardado en:
| Autores principales: | , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | https://arxiv.org/abs/2504.19191 |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866909595689025536 |
|---|---|
| author | Xiao, Liu Zhiyuan, Li Yueyu, Lin |
| author_facet | Xiao, Liu Zhiyuan, Li Yueyu, Lin |
| contents | The WuNeng architecture introduces a novel approach to enhancing the expressivity and power of large language models by integrating recurrent neural network (RNN)-based RWKV-7 with advanced attention mechanisms, prioritizing heightened contextual coherence over reducing KV cache size. Building upon the hybrid-head concept from Hymba, WuNeng augments standard multi-head attention with additional RWKV-7 state-driven heads, rather than replacing existing heads, to enrich the model's representational capacity. A cross-head interaction technique fosters dynamic synergy among standard, state-driven, and newly introduced middle heads, leveraging concatenation, additive modulation, and gated fusion for robust information integration. Furthermore, a multi-token state processing mechanism harnesses the continuous RWKV-7 state to capture intricate, sequence-wide dependencies, significantly boosting expressivity. Remarkably, these enhancements are achieved with minimal additional parameters, ensuring efficiency while empowering the model to excel in complex reasoning and sequence generation tasks. WuNeng sets a new standard for balancing expressivity and computational efficiency in modern neural architectures. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2504_19191 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | WuNeng: Hybrid State with Attention Xiao, Liu Zhiyuan, Li Yueyu, Lin Computation and Language The WuNeng architecture introduces a novel approach to enhancing the expressivity and power of large language models by integrating recurrent neural network (RNN)-based RWKV-7 with advanced attention mechanisms, prioritizing heightened contextual coherence over reducing KV cache size. Building upon the hybrid-head concept from Hymba, WuNeng augments standard multi-head attention with additional RWKV-7 state-driven heads, rather than replacing existing heads, to enrich the model's representational capacity. A cross-head interaction technique fosters dynamic synergy among standard, state-driven, and newly introduced middle heads, leveraging concatenation, additive modulation, and gated fusion for robust information integration. Furthermore, a multi-token state processing mechanism harnesses the continuous RWKV-7 state to capture intricate, sequence-wide dependencies, significantly boosting expressivity. Remarkably, these enhancements are achieved with minimal additional parameters, ensuring efficiency while empowering the model to excel in complex reasoning and sequence generation tasks. WuNeng sets a new standard for balancing expressivity and computational efficiency in modern neural architectures. |
| title | WuNeng: Hybrid State with Attention |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2504.19191 |