English
Related papers

Related papers: Test-Time Speculation

200 papers

Speculative sampling (SpS) has been successful in accelerating the decoding throughput of auto-regressive large language models by leveraging smaller draft models. SpS strictly enforces the generated distribution to match that of the…

Machine Learning · Computer Science 2026-04-08 Yongchang Hao , Lili Mou

Speculative decoding accelerates LLM inference by using a smaller draft model to speculate tokens that a larger target model verifies. Verification is often the bottleneck (e.g. verification is $4\times$ slower than token generation when a…

Computation and Language · Computer Science 2026-05-27 Avinash Kumar , Sujay Sanghavi , Poulami Das

The autoregressive nature of large language models (LLMs) fundamentally limits inference speed, as each forward pass generates only a single token and is often bottlenecked by memory bandwidth. Speculative decoding has emerged as a…

Machine Learning · Computer Science 2025-12-02 Zihao An , Huajun Bai , Ziqiong Liu , Dong Li , Emad Barsoum

The recent advancements in large language models (LLMs) have been extraordinary, yet the escalating inference costs associated with them present challenges in real-world applications. To address these challenges, we propose a novel approach…

Computation and Language · Computer Science 2024-06-07 Jiahao Liu , Qifan Wang , Jingang Wang , Xunliang Cai

To mitigate the high inference latency stemming from autoregressive decoding in Large Language Models (LLMs), Speculative Decoding has emerged as a novel decoding paradigm for LLM inference. In each decoding step, this method first drafts…

Computation and Language · Computer Science 2024-06-05 Heming Xia , Zhe Yang , Qingxiu Dong , Peiyi Wang , Yongqi Li , Tao Ge , Tianyu Liu , Wenjie Li , Zhifang Sui

Speculative decoding promises faster inference for large language models (LLMs), yet existing methods fail to generalize to real-world settings. Benchmarks typically assume short contexts (e.g., 2K tokens), whereas practical workloads…

Computation and Language · Computer Science 2025-10-10 Jaeseong Lee , seung-won hwang , Aurick Qiao , Gabriele Oliaro , Ye Wang , Samyam Rajbhandari

Speculative decoding is a widely used method that accelerates the generation process of large language models (LLMs) with no compromise in model performance. It achieves this goal by using an existing smaller model for drafting and then…

Computation and Language · Computer Science 2024-10-16 Weilin Zhao , Yuxiang Huang , Xu Han , Wang Xu , Chaojun Xiao , Xinrong Zhang , Yewei Fang , Kaihuo Zhang , Zhiyuan Liu , Maosong Sun

This tutorial presents a comprehensive introduction to Speculative Decoding (SD), an advanced technique for LLM inference acceleration that has garnered significant research interest in recent years. SD is introduced as an innovative…

Computation and Language · Computer Science 2025-03-04 Heming Xia , Cunxiao Du , Yongqi Li , Qian Liu , Wenjie Li

Speculative decoding accelerates inference for Large Language Models by using a lightweight draft model to propose candidate tokens that are verified in parallel by a larger target model. Prior work shows that the draft model often…

Computation and Language · Computer Science 2026-03-06 Ofir Ben Shoham

Speculative decoding (SD), where a small draft model is employed to propose draft tokens in advance and then the target model validates them in parallel, has emerged as a promising technique for LLM inference acceleration. Many endeavors to…

Computation and Language · Computer Science 2026-04-30 Tianyu Liu , Qitan Lv , Hao Li , Xing Gao , Xiao Sun , Xiaoyan Sun

The immense model sizes of large language models (LLMs) challenge deployment on memory-limited consumer GPUs. Although model compression and parameter offloading are common strategies to address memory limitations, compression can degrade…

Computation and Language · Computer Science 2025-10-10 Pei-Shuo Wang , Jian-Jia Chen , Chun-Che Yang , Chi-Chih Chang , Ning-Chi Huang , Mohamed S. Abdelfattah , Kai-Chiang Wu

Auto-regressive decoding in Large Language Models (LLMs) is inherently memory-bound: every generation step requires loading the model weights and intermediate results from memory (e.g., High-Bandwidth Memory (HBM) for GPU servers), making…

Machine Learning · Computer Science 2026-05-13 Yuning Han , Yangchenchen Jin , Dylan Zhao , Jingwei Sun

Speculative decoding accelerates LLM inference by utilizing otherwise idle computational resources during memory-to-chip data transfer. Current speculative decoding methods typically assume a considerable amount of available computing…

Computation and Language · Computer Science 2025-11-26 Luohe Shi , Zuchao Li , Lefei Zhang , Baoyuan Qi , Guoming Liu , Hai Zhao

Speculative decoding accelerates LLM inference by letting a small drafter propose multiple tokens which a large target model verifies once per speculation step. As vocabularies scale past 10e5 tokens,verification cost in the target model is…

Computation and Language · Computer Science 2026-02-04 Jinbin Zhang , Nasib Ullah , Erik Schultheis , Rohit Babbar

Long-context large language model (LLM) inference has become the norm for today's AI applications. However, it is severely bottlenecked by the increasing memory demands of its KV cache. Previous works have shown that self-speculative…

Machine Learning · Computer Science 2026-02-10 Yikang Yue , Yuqi Xue , Jian Huang

Test-time scaling (TTS) has emerged as a new frontier for scaling the performance of Large Language Models. In test-time scaling, by using more computational resources during inference, LLMs can improve their reasoning process and task…

Computation and Language · Computer Science 2025-09-10 V Venktesh , Mandeep Rathee , Avishek Anand

The sequential nature of modern LLMs makes them expensive and slow, and speculative sampling has proven to be an effective solution to this problem. Methods like EAGLE perform autoregression at the feature level, reusing top-layer features…

Computation and Language · Computer Science 2025-04-24 Yuhui Li , Fangyun Wei , Chao Zhang , Hongyang Zhang

With the increasingly giant scales of (causal) large language models (LLMs), the inference efficiency comes as one of the core concerns along the improved performance. In contrast to the memory footprint, the latency bottleneck seems to be…

Computation and Language · Computer Science 2024-04-24 Chen Zhang , Zhuorui Liu , Dawei Song

Large language models and large multimodal models (LLMs and LMMs) deliver strong generative performance but suffer from slow decoding, a problem that becomes more severe when handling visual inputs, whose sequences typically contain many…

Computer Vision and Pattern Recognition · Computer Science 2026-02-04 Zihua Wang , Ruibo Li , Haozhe Du , Joey Tianyi Zhou , Yu Zhang , Xu Yang

Conventional speculative decoding (SD) methods utilize a predefined length policy for proposing drafts, which implies the premise that the target model smoothly accepts the proposed draft tokens. However, reality deviates from this…

Computation and Language · Computer Science 2025-08-26 Ziyin Zhang , Jiahao Xu , Tian Liang , Xingyu Chen , Zhiwei He , Rui Wang , Zhaopeng Tu