Image de couverture de livre

Think Before You Accept: Semantic Reflective Verification for Faster Speculative Decoding

Fuente: arXiv

Enregistré dans:

Détails bibliographiques
Auteurs principaux:	Wang, Yixuan, Liu, Yijun, ji, Shiyu, Xu, Yuzhuang, Xu, Yang, Zhu, Qingfu, Che, Wanxiang
Format:	Preprint
Publié:	2025
Sujets:	Machine Learning
Accès en ligne:	Acceder al recurso
Tags:	Ajouter un tag Pas de tags, Soyez le premier à ajouter un tag!

Documents similaires

CRVQ: Channel-Relaxed Vector Quantization for Extreme Compression of LLMs
par: Xu, Yuzhuang, et autres
Publié: (2024)

CAMERA: Multi-Matrix Joint Compression for MoE Models via Micro-Expert Redundancy Analysis
par: Xu, Yuzhuang, et autres
Publié: (2025)

Lookahead Q-Cache: Achieving More Consistent KV Cache Eviction via Pseudo Query
par: Wang, Yixuan, et autres
Publié: (2025)

Judge Q: Trainable Queries for Optimized Information Retention in KV Cache Eviction
par: Liu, Yijun, et autres
Publié: (2025)

Make Some Noise: Unlocking Language Model Parallel Inference Capability through Noisy Training
par: Wang, Yixuan, et autres
Publié: (2024)

CommonKV: Compressing KV Cache with Cross-layer Parameter Sharing
par: Wang, Yixuan, et autres
Publié: (2025)

Seer Self-Consistency: Advance Budget Estimation for Adaptive Test-Time Scaling
par: Ji, Shiyu, et autres
Publié: (2025)

EchoKV: Efficient KV Cache Compression via Similarity-Based Reconstruction
par: Ji, Shiyu, et autres
Publié: (2026)

Greedy Multi-Path Block Verification for Faster Decoding in Speculative Sampling
par: Thomas, Rahul, et autres
Publié: (2026)

Multi-Layer Attention is the Amplifier of Demonstration Effectiveness
par: Wang, Dingzirui, et autres
Publié: (2025)

ProxyAttn: Guided Sparse Attention via Representative Heads
par: Wang, Yixuan, et autres
Publié: (2025)

Think Before You Prune: Self-Reflective Structured Pruning for Reasoning Language Models
par: Wang, Ziyan, et autres
Publié: (2025)

Faster Cascades via Speculative Decoding
par: Narasimhan, Harikrishna, et autres
Publié: (2024)

Semantic-Guided Generative Image Augmentation Method with Diffusion Models for Image Classification
par: Li, Bohan, et autres
Publié: (2023)

Self-Speculative Biased Decoding for Faster Re-Translation
par: Zeng, Linxiao, et autres
Publié: (2025)

Traversal Verification for Speculative Tree Decoding
par: Weng, Yepeng, et autres
Publié: (2025)

MARS: Unleashing the Power of Speculative Decoding via Margin-Aware Verification
par: Song, Jingwei, et autres
Publié: (2026)

DySpec: Faster Speculative Decoding with Dynamic Token Tree Structure
par: Xiong, Yunfan, et autres
Publié: (2024)

TriSpec: Ternary Speculative Decoding via Lightweight Proxy Verification
par: Jiang, Haoyun, et autres
Publié: (2026)

Think Twice Before You Act: Improving Inverse Problem Solving With MCMC
par: Zhu, Yaxuan, et autres
Publié: (2024)

HUOZIIME: An On-Device LLM-enhanced Input Method for Deep Personalization
par: Shan, Baocai, et autres
Publié: (2026)

Speculative Speculative Decoding
par: Kumar, Tanishq, et autres
Publié: (2026)

Block Verification Accelerates Speculative Decoding
par: Sun, Ziteng, et autres
Publié: (2024)

OneBit: Towards Extremely Low-bit Large Language Models
par: Xu, Yuzhuang, et autres
Publié: (2024)

Speculative Safety-Aware Decoding
par: Wang, Xuekang, et autres
Publié: (2025)

Judge Decoding: Faster Speculative Sampling Requires Going Beyond Model Alignment
par: Bachmann, Gregor, et autres
Publié: (2025)

Speeding up Speculative Decoding via Sequential Approximate Verification
par: Zhong, Meiyu, et autres
Publié: (2025)

Think Before You Act: Decision Transformers with Working Memory
par: Kang, Jikun, et autres
Publié: (2023)

Balancing Coverage and Draft Latency in Vocabulary Trimming for Faster Speculative Decoding
par: Shoham, Ofir Ben
Publié: (2026)

Annealed Relaxation of Speculative Decoding for Faster Autoregressive Image Generation
par: Li, Xingyao, et autres
Publié: (2026)

Improving Grammatical Error Correction via Contextual Data Augmentation
par: Wang, Yixuan, et autres
Publié: (2024)

Speculative Decoding for Verilog: Speed and Quality, All in One
par: Xu, Changran, et autres
Publié: (2025)

LongSpec: Long-Context Lossless Speculative Decoding with Efficient Drafting and Verification
par: Yang, Penghui, et autres
Publié: (2025)

Speculative Decoding Meets Quantization: Compatibility Evaluation and Hierarchical Framework Design
par: Zhang, Yudi, et autres
Publié: (2025)

Decoding Speculative Decoding
par: Yan, Minghao, et autres
Publié: (2024)

Think Before You Lie: How Reasoning Leads to Honesty
par: Yuan, Ann, et autres
Publié: (2026)

Dynamic Adaptive Shared Experts with Grouped Multi-Head Attention Mixture of Experts
par: Li, Cheng, et autres
Publié: (2025)

Advancing Tool-Augmented Large Language Models via Meta-Verification and Reflection Learning
par: Ma, Zhiyuan, et autres
Publié: (2025)

ArcLight: A Lightweight LLM Inference Architecture for Many-Core CPUs
par: Xu, Yuzhuang, et autres
Publié: (2026)

SpecPV: Improving Self-Speculative Decoding for Long-Context Generation via Partial Verification
par: Tan, Zhendong, et autres
Publié: (2025)