SpecPV: Improving Self-Speculative Decoding for Long-Context Generation via Partial Verification

arXiv:2512.02337 · cs.LG · Submitted 2025-12-02 · Read on arXiv

cs.LG

Submitted: 2025-12-02

Updated: 2026-08-29

Code: https://github.com/TanZhendong/SpecPV

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Terminology

• Verification — 88× in this paper · explained in 2 episodes
• KV cache — 54× in this paper · explained in 3 episodes
• Speculative Decoding — 46× in this paper · explained in 2 episodes
• Accuracy — 13× in this paper · explained in 1 episode
• Benchmark — 4× in this paper · explained in 2 episodes
• Quantization — 4× in this paper · explained in 12 episodes
• Sparse Attention — 3× in this paper · explained in 5 episodes
• Sparsity — 3× in this paper · explained in 4 episodes
• Adaptation — 2× in this paper · explained in 1 episode
• Fine-Tuning — 2× in this paper · explained in 5 episodes
• Large Language Models (LLMs) — 2× in this paper · explained in 5 episodes
• Transformer — 2× in this paper · explained in 2 episodes
• Ablation — 1× in this paper · explained in 1 episode
• Ablation study — 1× in this paper · explained in 1 episode
• Aggregation — 1× in this paper · explained in 1 episode
• Consistency — 1× in this paper · explained in 3 episodes
• Core — 1× in this paper · explained in 2 episodes
• Embeddings — 1× in this paper · explained in 2 episodes
• Exponential Moving Average (EMA) — 1× in this paper · explained in 2 episodes
• Semantic Similarity — 1× in this paper · explained in 2 episodes

Sources

• DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning
• OpenAI o1 System Card
• LongBench v2: Towards Deeper Understanding and Reasoning on Realistic Long-context Multitasks
• EAGLE-3: Scaling up Inference Acceleration of Large Language Models via Training-Time Test
• The Llama 3 Herd of Models
• DeepSeek-V3 Technical Report
• DeepSeek-OCR: Contexts Optical Compression
• YaRN: Efficient Context Window Extension of Large Language Models
• Qwen3 Technical Report
• Gated Delta Networks: Improving Mamba2 with Delta Rule
• LServe: Efficient Long-sequence LLM Serving with Unified Sparse Attention
• Quest: Query-Aware Sparsity for Efficient Long-Context LLM Inference
• Kimi Linear: An Expressive, Efficient Attention Architecture

Related papers

• Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
• AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
• Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
• Convergence issues in Relational Concept Analysis based on AOC-posets
• Beliefs Beyond Posteriors: Local-Consistency Optimisation for Bayesian Neural Networks
• Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks