SharpDraft: Accelerating Long-Context Speculative Decoding with Cardinality-Aware Query Scaling
cs.AI
Submitted: 2026-10-04
Updated: 2026-10-04
Project page: https://vllm-project.github.io/2026/05/26/eagle-3-1.html
Terminology
Sources
- Accelerating Large Language Model Decoding with Speculative Sampling
- Beyond Benchmarks: MathArena as an Evaluation Platform for Mathematics with LLMs
- Attention Drift: What Autoregressive Speculative Decoding Models Learn
- AgentLongBench: A Controllable Long Benchmark For Long-Contexts Agents via Environment Rollouts
- Test-Time Speculation
- LongReason: A Synthetic Long-Context Reasoning Benchmark via Context Expansion
- LongCoT: Benchmarking Long-Horizon Chain-of-Thought Reasoning
- Scalable-Softmax Is Superior for Attention
- LongGenBench: Benchmarking Long-Form Generation in Long Context LLMs
Related papers
- MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
- Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
- Agents' Last Exam
- The Clinician's Veto: Navigating Trust, Liability, and Uncertainty in Autonomous AI Prescribing
- MindHelper: Closed-Loop Embodied Mental-State Reasoning for Precision Intervention
- Incumbent Advantage: Brand Bias and Cognitive Manipulation Dynamics in LLM Recommendation Systems