BMRUs enable analog recurrent neural network hardware via discrete outputs that suppress noise 20-fold, with one-to-one parameter-to-circuit mapping and linear power scaling for recurrence.
The end of transformers? On challenging attention and the rise of sub-quadratic architectures
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
years
2026 3representative citing papers
xLSTM outperforms Mamba-2 and Gated DeltaNet on tasks with complex dependencies because its gating scheme enables more flexible and stable state tracking and memory accumulation.
Reasoning-token augmentation dominates architectural bias for state-based recall tasks; hybrid advantages are narrow and task-dependent rather than uniform.
citing papers explorer
-
Hardware-Software Co-Design of Scalable, Energy-Efficient Analog Recurrent Computations
BMRUs enable analog recurrent neural network hardware via discrete outputs that suppress noise 20-fold, with one-to-one parameter-to-circuit mapping and linear power scaling for recurrence.
-
On Subquadratic Architectures: From Applications to Principles
xLSTM outperforms Mamba-2 and Gated DeltaNet on tasks with complex dependencies because its gating scheme enables more flexible and stable state tracking and memory accumulation.
-
Reasoning Primitives in Hybrid and Non-Hybrid LLMs: Do Architectural Differences Yield Advantages in State-Tracking and Recall?
Reasoning-token augmentation dominates architectural bias for state-based recall tasks; hybrid advantages are narrow and task-dependent rather than uniform.