analiza aiqbench
This arXiv preprint proposes a speculative decoding method that lets draft tokens query the target model's recurrent states, aimed at speeding up LLM inference on hybrid linear attention architectures. The reported 3.16x average speedup and roughly 8% gains over DSpark are self-reported on seven benchmarks, so they need independent replication before they can be taken as settled. For readers, the useful signal is that recurrent state is becoming a reusable resource for inference efficiency, not just a model internal.
ReaDy: Recurrent States as Dynamic Context for Speculative Drafting in Hybrid Linear Attention Large Language Models
arXiv:2610.20503v1 Announce Type: new Abstract: Speculative decoding accelerates autoregressive generation by drafting multiple tokens per step, yet feature-based drafters rely on target hidden features whose contextual emphasis may not align with the needs of subsequent draft positions. Hybrid linear attention (HLA) models maintain compact recurrent states that accumulate historical associations, offering an efficient source of complementary context for speculative drafting. We introduce ReaDy, the first speculative decoding method, to the best of our knowledge, to leverage HLA recurrent states for position-specific context retrieval during drafting. ReaDy equips each draft position with a learned query to retrieve supplementary context directly from the target's committed recurrent states. A lightweight state reader integrates this context into draft representations through residual updates, while low-rank state reads and fused operators keep additional computation and I/O overhead low. Building on DSpark's parallel backbone and sequential prediction head, ReaDy employs a lightweight two-stage adaptation strategy, with reader capacity, state sources, and residual placement examined through ablation studies. Across seven benchmarks spanning mathematics, code generation, and dialogue, ReaDy achieves an average $3.16\times$ decoding speedup over autoregressive decoding with Qwen3.8-27B on a single NVIDIA B200 GPU. It outperforms DSpark on every benchmark, improving throughput by 7.97% and acceptance length by 8.50% on average. Our code is available at https://github.com/StartLuxLabs/ReaDy.
Abstract:Speculative decoding accelerates autoregressive generation by drafting multiple tokens per step, yet feature-based drafters rely on target hidden features whose contextual emphasis may not align with the needs of subsequent draft positions. Hybrid linear attention (HLA) models maintain compact recurrent states that accumulate historical associations, offering an efficient source of complementary context for speculative drafting. We introduce ReaDy, the first speculative decoding method, to the best of our knowledge, to leverage HLA recurrent states for position-specific context retrieval during drafting. ReaDy equips each draft position with a learned query to retrieve supplementary context directly from the target's committed recurrent states. A lightweight state reader integrates this context into draft representations through residual updates, while low-rank state reads and fused operators keep additional computation and I/O overhead low. Building on DSpark's parallel backbone and sequential prediction head, ReaDy employs a lightweight two-stage adaptation strategy, with reader capacity, state sources, and residual placement examined through ablation studies. Across seven benchmarks spanning mathematics, code generation, and dialogue, ReaDy achieves an average $3.16\times$ decoding speedup over autoregressive decoding with Qwen3.8-27B on a single NVIDIA B200 GPU. It outperforms DSpark on every benchmark, improving throughput by 7.97% and acceptance length by 8.50% on average. Our code is available at this https URL.
Computation and Language (cs.CL)
arXiv:2610.20503 [cs.CL]
(or arXiv:2610.20503v1 [cs.CL] for this version)
https://doi.org/10.48550/arXiv.2610.20503 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 302 cuvinte
Subiecte conexe
Deploying Qwen3.8-2.4T-A95B on Amazon SageMaker HyperPod with vLLM
AWS · 9 sept. 2026
[AINews] Muse Spark 1.3 matches GPT-5.6-Sol, confirming Meta Superintelligence as the newest Frontier Lab, >90% discount for training
Latent Space · 3 sept. 2026
ReplaySSM: Rethinking State Storage for Faster Hybrid-Model Decoding
NVIDIA · 12 oct. 2026
[AINews] not much happened today
Latent Space · 3 oct. 2026