arXiv AI
Jul 7
Training Hybrid Block Diffusion Language Models with Partial Bidirectionality
arXiv:2607. 02805v1 Announce Type: cross Abstract: High-throughput long-context generation is one of the central challenges for large language models.
Fixed State, Long Reach: What a Constant-Size Cache Buys Block Diffusion at Scale
The paper investigates how block‑diffusion language models can use a constant‑size cache to enable efficient parallel decoding. By employing sequence mixers that summarize completed blocks into a reusable state and a block‑causal training objective, the authors pretrain three 3B block‑diffusion denoisers (attention, Mamba, and hybrid) on 300 B tokens. The resulting state‑space cache remains O(1) in memory and latency regardless of context length, yielding significant speed‑up and memory savings compared to traditional attention‑based caches, especially at very long sequences.
Related stories
arXiv AI
Sep 1
DASC: Decay-Aware State Compression for Hybrid Linear-Attention Serving
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...
arXiv Machine Learning
Aug 10
Retrofitting Linear Attention into Diffusion Language Models
arXiv:2608. 06628v1 Announce Type: new Abstract: Diffusion language models (dLLMs) offer a promising alternative to autoregressive models by accelerating inference through parallel decoding.
arXiv AI
Oct 15
GateSight: Recurrent Memory Anticipates Attention Memory for Query-Agnostic KV Eviction in Hybrid LLMs
arXiv:2610.14754v1 Announce Type: new Abstract: Hybrid language models interleave Gated DeltaNet (GDN) layers, whose recurrent memory has a fixed size, with full-attention layers, whose key-value (KV...
arXiv AI
Jul 21
FlowBlock: Wavefront-Parallel Decoding for Self-Correcting Diffusion Language Models
arXiv:2607. 17652v1 Announce Type: new Abstract: Block-wise diffusion large language models (dLLMs) decode sequentially at the block level, enabling effective KV-cache reuse across blocks but making inter-block decoding strictly serial.
arXiv AI
Jun 29
Bifocal Diffusion Language Models: Asymmetric Bidirectional Context for Parallel Generation
arXiv:2606. 27732v1 Announce Type: cross Abstract: Discrete diffusion language models (dLLMs) recover masked tokens in parallel, offering significant speedups over autoregressive (AR) generation.
Hugging Face Trending Papers
Jul 5
Sangam: Efficiently Serving Diffusion LLMs with the AR Stack
Diffusion language models (dLLMs) generate text by iteratively denoising a masked response and can commit multiple output positions per model invocation. Their bidirectional attention prevents exact autoregressive-style KV caching, since committing one position shifts the KV activations of all others.
Hugging Face Trending Papers
Jul 20
FlowBlock: Wavefront-Parallel Decoding for Self-Correcting Diffusion Language Models
Block-wise diffusion large language models (dLLMs) decode sequentially at the block level, enabling effective KV-cache reuse across blocks but making inter-block decoding strictly serial. Prior work has attempted to unlock inter-block parallelism through post-training methods, but achieves only modest speedups and often degrades accuracy.
arXiv AI
Aug 13
LinearKV: One Cached State Suffices for Position-Independent Caching in Hybrid LLMs
arXiv:2608. 11231v1 Announce Type: new Abstract: LLM serving is increasingly accelerated by position-independent caching (PIC).
arXiv Machine Learning
Sep 24
DeltaS: Reading the Gated Linear Attention State for KV Cache Eviction in Streaming Video
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.
arXiv AI
Aug 21
Daedalus-150M: A Convolution-Attention Hybrid Designed for CPU Inference
arXiv:2608. 20210v1 Announce Type: cross Abstract: Small language models are usually built like large ones and then squeezed onto a CPU afterwards.
arXiv AI
Oct 7
Mask-Guided KV Cache Eviction in Block Diffusion Language Models
arXiv:2610.06996v1 Announce Type: cross Abstract: Block diffusion language models keep a large key-value (KV) cache throughout generation and attend to it at every denoising step, limiting both memor...