SwiGLU 技术
领域: Artificial Intelligence
In neural networks, the gating mechanism is an architectural motif for controlling the flow of activation and gradient signals. They are most prominently used in recurrent neural networks (RNNs), but have also found applications in other architectures.
2
衍生技术
0
相关产品
7
相关事件
相关专利
暂无数据
相关产品
暂无数据
相关事件查看全部 (7)
Confidence-Adaptive SwiGLU for Mixture-of-Experts
2026-06-02PRODUCT_LAUNCH影响: MEDIUM
CosmicFish-HRM: Adaptive Reasoning via Hierarchical Recurrent Mechanisms in Compact Language Models
2026-05-29PRODUCT_LAUNCH影响: MEDIUM
More Expressive Feedforward Layers: Part I. Token-Adaptive Mixing of Activations
2026-05-27PRODUCT_LAUNCH影响: MEDIUM
Symmetry-Compatible Principle for Optimizer Design: Embeddings, LM Heads, SwiGLU MLPs, and MoE Routers
2026-05-27PRODUCT_LAUNCH影响: MEDIUM
相关文章查看全部 (7)
Measuring Maximum Activations in Open Large Language Models
ArXiv CS.CL2026-06-03
Confidence-Adaptive SwiGLU for Mixture-of-Experts
ArXiv CS.CL2026-06-02
More Expressive Feedforward Layers: Part I. Token-Adaptive Mixing of Activations
ArXiv CS.AI2026-05-27