Mistletoe: Stealthy Acceleration-Collapse Attacks on Speculative Decoding

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Sun, Shuoyang, Dai, Chang, Fang, Hao, Gao, Kuofeng, Zhong, Xinhao, Sun, Yi, Mo, Fan, Xia, Shu-Tao, Chen, Bin
Format:	Preprint
Published:	2026
Subjects:	Computation and Language Machine Learning
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Revisiting Backdoor Attacks on LLMs: A Stealthy and Practical Poisoning Framework via Harmless Inputs
by: Kong, Jiawei, et al.
Published: (2025)

Grounding Language with Vision: A Conditional Mutual Information Calibrated Decoding Strategy for Reducing Hallucinations in LVLMs
by: Fang, Hao, et al.
Published: (2025)

LogitSpec: Accelerating Retrieval-based Speculative Decoding via Next Next Token Speculation
by: Liu, Tianyu, et al.
Published: (2025)

Your Language Model Can Secretly Write Like Humans: Contrastive Paraphrase Attacks on LLM-Generated Text Detectors
by: Fang, Hao, et al.
Published: (2025)

SpecHub: Provable Acceleration to Multi-Draft Speculative Decoding
by: Sun, Ryan, et al.
Published: (2024)

Denial-of-Service Poisoning Attacks against Large Language Models
by: Gao, Kuofeng, et al.
Published: (2024)

Accelerate Speculative Decoding with Sparse Computation in Verification
by: Wang, Jikai, et al.
Published: (2025)

Differential Vector Erasure: Unified Training-Free Concept Erasure for Flow Matching Models
by: Zhang, Zhiqi, et al.
Published: (2026)

Block Verification Accelerates Speculative Decoding
by: Sun, Ziteng, et al.
Published: (2024)

SWIFT: On-the-Fly Self-Speculative Decoding for LLM Inference Acceleration
by: Xia, Heming, et al.
Published: (2024)

Prompt2Fingerprint: Plug-and-Play LLM Fingerprinting via Text-to-Weight Generation
by: Chen, Sixu, et al.
Published: (2026)

Towards Distillation-Resistant Large Language Models: An Information-Theoretic Perspective
by: Fang, Hao, et al.
Published: (2026)

Dynamic Speculation Lookahead Accelerates Speculative Decoding of Large Language Models
by: Mamou, Jonathan, et al.
Published: (2024)

EAGLE-Pangu: Accelerator-Safe Tree Speculative Decoding on Ascend NPUs
by: Han, Chang, et al.
Published: (2026)

TriForce: Lossless Acceleration of Long Sequence Generation with Hierarchical Speculative Decoding
by: Sun, Hanshi, et al.
Published: (2024)

Speculative Contrastive Decoding
by: Yuan, Hongyi, et al.
Published: (2023)

Multi-Candidate Speculative Decoding
by: Yang, Sen, et al.
Published: (2024)

FlowErase-RL: Rethinking Concept Erasure as Reward Optimization in Flow Matching Models
by: Sun, Yi, et al.
Published: (2026)

Speculate Deep and Accurate: Lossless and Training-Free Acceleration for Offloaded LLMs via Substitute Speculative Decoding
by: Wang, Pei-Shuo, et al.
Published: (2025)

TALON: Confidence-Aware Speculative Decoding with Adaptive Token Trees
by: Liu, Tianyu, et al.
Published: (2026)

Stealthy Backdoor Attacks against LLMs Based on Natural Style Triggers
by: Wei, Jiali, et al.
Published: (2026)

Benchmarking Open-ended Audio Dialogue Understanding for Large Audio-Language Models
by: Gao, Kuofeng, et al.
Published: (2024)

Accelerating Speculative Decoding with Block Diffusion Draft Trees
by: Ringel, Liran, et al.
Published: (2026)

Efficient Adaptive Rejection Sampling for Accelerating Speculative Decoding in Large Language Models
by: Sun, Chendong, et al.
Published: (2025)

Large Language Models are Good Attackers: Efficient and Stealthy Textual Backdoor Attacks
by: Li, Ziqiang, et al.
Published: (2024)

ViSpec: Accelerating Vision-Language Models with Vision-Aware Speculative Decoding
by: Kang, Jialiang, et al.
Published: (2025)

GaussTrap: Stealthy Poisoning Attacks on 3D Gaussian Splatting for Targeted Scene Confusion
by: Hong, Jiaxin, et al.
Published: (2025)

Cacheback: Speculative Decoding With Nothing But Cache
by: Ma, Zhiyao, et al.
Published: (2025)

ToolSpec: Accelerating Tool Calling via Schema-Aware and Retrieval-Augmented Speculative Decoding
by: Xia, Heming, et al.
Published: (2026)

Cross-Attention Speculative Decoding
by: Zhong, Wei, et al.
Published: (2025)

Goose: Anisotropic Speculation Trees for Training-Free Speculative Decoding
by: Jin, Tao, et al.
Published: (2026)

COLD-Attack: Jailbreaking LLMs with Stealthiness and Controllability
by: Guo, Xingang, et al.
Published: (2024)

Clover: Regressive Lightweight Speculative Decoding with Sequential Knowledge
by: Xiao, Bin, et al.
Published: (2024)

Scaling Laws for Speculative Decoding
by: Yan, Siyuan, et al.
Published: (2025)

Speculative Decoding with a Speculative Vocabulary
by: Williams, Miles, et al.
Published: (2026)

Decoding Speculative Decoding
by: Yan, Minghao, et al.
Published: (2024)

RASD: Retrieval-Augmented Speculative Decoding
by: Quan, Guofeng, et al.
Published: (2025)

Cost-Aware Diffusion Draft Trees for Speculative Decoding
by: Zhang, Shuai, et al.
Published: (2026)

Online Speculative Decoding
by: Liu, Xiaoxuan, et al.
Published: (2023)

AdaServe: Accelerating Multi-SLO LLM Serving with SLO-Customized Speculative Decoding
by: Li, Zikun, et al.
Published: (2025)