arXiv:2608.30310v1 Announce Type: cross
Abstract: Hybrid large language models interleave full-attention layers with linear-attention layers to reduce the cost of long-context inference. This structu...
By Yirui Liu, Ruoling Qi, Xuaner Wu, Penghang Liu, Jian Chen
arXiv:2608.30386v1 Announce Type: cross
Abstract: Hybrid linear-attention architectures have recently scaled to large open-weight models, offering quality competitive with full attention while substa...
By Yanqi Yu, Pingwei Sun, Jianchao Tan, Tao Zhang, Yuchen Xie, Xunliang Cai, Yao Liu
arXiv:2608. 12435v1 Announce Type: new Abstract: Transformers owe much of their strong long-context retrieval capability to a token-level memory that grows with context length.
By Ming Zhang, Kaisen Yang, Shu Yu, Ermo Hua, Ning Ding, Xia Hu, Bowen Zhou, Chaochao Lu, Youbang Sun
arXiv:2606. 09888v1 Announce Type: new Abstract: Linear attention provides an efficient backbone for long-sequence recommendation by avoiding the quadratic cost of standard Transformers, but its compressed recurrent state can be dominated by repetitive behavior patterns.
By Zhuang Zhuang, Zhipeng Wei, Ji Dai, Jie Chen, Fei Pan, Peng Jiang, Kun Gai
DeltaS is a query‑agnostic, training‑free method for evicting key‑value cache entries in hybrid video‑language models that combine linear and full attention. It uses the change in the recurrent state of gated‑delta linear attention—called state drift—to decide which video chunks to keep, selecting those that induce larger normalized state changes. In experiments with a fixed memory budget, DeltaS outperforms position‑, attention‑, and key‑value‑based eviction signals, improving performance by 2.1 points on average across six long‑video benchmarks and 5.6 points on the longest benchmark, while adding only 1.9% of the forward‑pass cost.
By Taeyoun Kwon, Seungjin Kim, Hyeonyu Kim, Moon Hwan Kim
Elastic Threshold Attention (ETA) is a trainable attention mechanism that dynamically predicts contextual thresholds from query representations, enabling selective pruning of KV cache tokens during long‑context decoding. By multiplicatively suppressing sub‑threshold logits during training, ETA avoids representation collapse and eliminates localized attention sinks, allowing a 1.45B model to match dense attention performance at roughly 85% training sparsity and 38% active decode density. At inference, a custom Triton kernel achieves up to 2.5× faster decoding on sequences up to 512K tokens, and an offline calibration step can further reduce compute by 27% by freezing per‑head thresholds.
By Themistoklis Haris, Henry Li, Maryam Karimzadehgan