arXiv:2606. 15678v1 Announce Type: cross Abstract: A feasibility and dynamics study of the Reservoir Attention Network (RAN), an architecture that injects a fixed, randomly-initialized reservoir into the mid-layer attention of a pretrained transformer to carry state across forward passes.
arXiv:2605.26797v2 Announce Type: replace Abstract: We study Latent Recurrent Transformer (LRT), a lightweight augmentation of autoregressive transformers that reuses a high-level source-layer hidden...
arXiv:2601. 11667v2 Announce Type: replace-cross Abstract: Transformer architectures deliver state-of-the-art accuracy via dense full-attention, but their quadratic time and memory complexity with respect to sequence length limits practical deployment.
arXiv:2607. 22696v1 Announce Type: cross Abstract: High-resolution video diffusion models built on Diffusion Transformers (DiTs) deliver strong fidelity but quickly exhaust the memory budget of a single workstation.
arXiv:2606. 06453v1 Announce Type: new Abstract: Sparse attention is becoming increasingly important for serving large language models (LLMs) as generation lengths continue to grow.
arXiv:2603. 06591v2 Announce Type: replace Abstract: Transformers frequently allocate disproportionate attention to specific tokens, a phenomenon known as attention sinks.
InfoMamba is an attention‑free hybrid model that combines a minimal‑bandwidth global interface with a selective recurrent stream. The architecture replaces token‑level self‑attention with a concept bottleneck linear filtering layer and integrates it via an information‑maximizing fusion (IMF) that injects global context into the state‑space dynamics. Experiments across classification, dense prediction, and non‑vision tasks show that InfoMamba outperforms strong Transformer and SSM baselines while maintaining near‑linear scaling and competitive accuracy‑efficiency trade‑offs.
arXiv:2607. 18910v1 Announce Type: new Abstract: Sequential decision making in non-stationary and partially observable environments requires rapid adaptation to latent regime changes.
arXiv:2502.09245v3 Announce Type: replace Abstract: In contrast to RNNs, which compress their history into a single hidden state, Transformers can attend to all past tokens directly. However, standar...
arXiv:2608. 02870v1 Announce Type: new Abstract: We introduce \ours{}, a recurrent Transformer architecture with fixed-size memory that generalizes sliding-window attention while remaining parallelizable during training.
arXiv:2608.30720v1 Announce Type: new Abstract: Representational similarity is foundational to analyses of deep networks, yet distances between point-valued representations are not intrinsically tied...
arXiv:2604. 10098v2 Announce Type: replace Abstract: As the foundational architecture of modern machine learning, Transformers have driven remarkable progress across diverse AI domains.
arXiv:2606. 16730v2 Announce Type: replace-cross Abstract: We re-interpret Transformer pretraining as a fast-slow, singularly perturbed flow along depth, with untied weights as its non-autonomous feature.