Cover Image

SpecExit: Accelerating Large Reasoning Model via Speculative Exit

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Yang, Rubing, Bai, Huajun, Liu, Song, Yu, Guanghua, Fan, Runzhi, Dang, Yanbin, Zhang, Jiejing, Liu, Kai, Zhu, Jianchen, Chen, Peng
Format:	Preprint
Published:	2025
Subjects:	Artificial Intelligence Computation and Language Machine Learning
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

NEAT: Neuron-Based Early Exit for Large Reasoning Models
by: Liu, Kang, et al.
Published: (2026)

Dynamic Early Exit in Reasoning Models
by: Yang, Chenxu, et al.
Published: (2025)

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

LayerSkip: Enabling Early Exit Inference and Self-Speculative Decoding
by: Elhoushi, Mostafa, et al.
Published: (2024)

Pipeline Parallelism is All You Need for Optimized Early-Exit Based Self-Speculative Decoding
by: Li, Ruanjun, et al.
Published: (2025)

Accelerating Large Language Model Inference with Self-Supervised Early Exits
by: Valade, Florian
Published: (2024)

DFlare: Scaling Up Draft Capacity for Block Diffusion Speculative Decoding
by: Zhang, Jiebin, et al.
Published: (2026)

DEL: Context-Aware Dynamic Exit Layer for Efficient Self-Speculative Decoding
by: Zarch, Hossein Entezari, et al.
Published: (2025)

FlashThink: An Early Exit Method For Efficient Reasoning
by: Jiang, Guochao, et al.
Published: (2025)

When Is Thinking Enough? Early Exit via Sufficiency Assessment for Efficient Reasoning
by: Xiang, Yang, et al.
Published: (2026)

ConfSpec: Efficient Step-Level Speculative Reasoning via Confidence-Gated Verification
by: Liu, Siran, et al.
Published: (2026)

FR-Spec: Accelerating Large-Vocabulary Language Models via Frequency-Ranked Speculative Sampling
by: Zhao, Weilin, et al.
Published: (2025)

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

Early Exit Is a Natural Capability in Transformer-based Models: An Empirical Study on Early Exit without Joint Optimization
by: Shan, Weiqiao, et al.
Published: (2024)

PEARL: Parallel Speculative Decoding with Adaptive Draft Length
by: Liu, Tianyu, et al.
Published: (2024)

LYNX: Learning Dynamic Exits for Confidence-Controlled Reasoning
by: Akgül, Ömer Faruk, et al.
Published: (2025)

Stop When Reasoning Converges: Semantic-Preserving Early Exit for Reasoning Models
by: Min, Dehai, et al.
Published: (2026)

Accelerating Multimodal Large Language Models via Dynamic Visual-Token Exit and the Empirical Findings
by: Wu, Qiong, et al.
Published: (2024)

HeteroSpec: Leveraging Contextual Heterogeneity for Efficient Speculative Decoding
by: Liu, Siran, et al.
Published: (2025)

BEEM: Boosting Performance of Early Exit DNNs using Multi-Exit Classifiers as Experts
by: Bajpai, Divya Jyoti, et al.
Published: (2025)

The Zero-Step Thinking: An Empirical Study of Mode Selection as Harder Early Exit in Reasoning Models
by: Tan, Yuqiao, et al.
Published: (2025)

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

NanoSpec: Accelerating Speculative Decoding using Minimalist In-Context Vocabularies
by: Chen, Zhiyang, et al.
Published: (2026)

SpecHop: Continuous Speculation for Accelerating Multi-Hop Retrieval Agents
by: Saberi, Mehrdad, et al.
Published: (2026)

SpecEyes: Accelerating Agentic Multimodal LLMs via Speculative Perception and Planning
by: Huang, Haoyu, et al.
Published: (2026)

SpecBlock: Block-Iterative Speculative Decoding with Dynamic Tree Drafting
by: Shi, Weijie, et al.
Published: (2026)

The Diminishing Returns of Early-Exit Decoding in Modern LLMs
by: Wei, Rui, et al.
Published: (2026)

SpecTr-GBV: Multi-Draft Block Verification Accelerating Speculative Decoding
by: Lin, Yijun, et al.
Published: (2026)

River-LLM: Large Language Model Seamless Exit Based on KV Share
by: Shen, Yingtao, et al.
Published: (2026)

Accelerating Large Language Model Reasoning via Speculative Search
by: Wang, Zhihai, et al.
Published: (2025)

Revisiting the Exit from Nuclear Energy in Germany with NLP
by: Haunss, Sebastian, et al.
Published: (2024)

Early-Exit and Instant Confidence Translation Quality Estimation
by: Zouhar, Vilém, et al.
Published: (2025)

TERMINATOR: Learning Optimal Exit Points for Early Stopping in Chain-of-Thought Reasoning
by: Nagle, Alliot, et al.
Published: (2026)

Dynamic Vocabulary Pruning in Early-Exit LLMs
by: Vincenti, Jort, et al.
Published: (2024)

SlimSpec: Low-Rank Draft LM-Head for Accelerated Speculative Decoding
by: Plaksin, Anton, et al.
Published: (2026)

SpecInfer: Accelerating Generative Large Language Model Serving with Tree-based Speculative Inference and Verification
by: Miao, Xupeng, et al.
Published: (2023)

Runaway is Ashamed, But Helpful: On the Early-Exit Behavior of Large Language Model-based Agents in Embodied Environments
by: Lu, Qingyu, et al.
Published: (2025)

BitSkip: An Empirical Analysis of Quantization and Early Exit Composition in Transformers
by: Bhuvaneswaran, Ramshankar, et al.
Published: (2025)

ADEPT: Adaptive Dynamic Early-Exit Process for Transformers
by: Yoo, Sangmin, et al.
Published: (2026)

CEEBERT: Cross-Domain Inference in Early Exit BERT
by: Bajpai, Divya Jyoti, et al.
Published: (2024)