SpecBlock achieves 8-13% higher mean speedup than EAGLE-3 at 44-52% drafting cost via block-iterative drafting with hidden-state inheritance, dynamic rank-head branching, valid-prefix masking, and optional cost-aware bandit adaptation.
hub
Fast inference from transformers via speculative decoding
10 Pith papers cite this work. Polarity classification is still indexing.
10
Pith papers citing it
hub tools
citation-role summary
background 2
method 1
citation-polarity summary
representative citing papers
DMax uses On-Policy Uniform Training and Soft Parallel Decoding to enable aggressive parallelism in dLLMs, raising TPF on GSM8K from 2.04 to 5.47 and on MBPP from 2.71 to 5.86 while preserving accuracy.
PPOW uses window-level RL with cost-aware speedup and proximity rewards plus adaptive divergence-aware windowing to reach 6.29-6.52 acceptance lengths and 3.39-4.36x speedups in speculative decoding.
A3 adaptively sets a VLA model's execution horizon to the longest action prefix the model re-confirms under conditional and sequential re-decoding, matching or improving task success without manual tuning.
Drafter models in speculative decoding suffer progressive attention drift caused by monotonically growing hidden-state magnitudes along the residual path; post-norm plus per-state RMSNorm reduces this drift and improves acceptance length up to 2x on perturbed templates and 1.18x on long-context data
Speculative Verification adds a companion model that estimates draft-target alignment via information gain to dynamically set verification length, delivering up to 2x speedup over standard speculative decoding across tested models and batch sizes.
SpecHop accelerates multi-hop LLM tool use via continuous multi-threaded speculation with asynchronous verification, approaching oracle latency gains and reducing latency up to 40% on retrieval tasks.
LAR learns a compact latent action space from trajectories that shortens the effective decision horizon for LLM agents, reducing token count and inference time while preserving task success.
Lever optimizes the drafting, verification, and execution stages of speculative decoding for flash-backed LLM inference on smartphones, reporting 2.93x average latency reduction over baseline flash-offloaded inference.
The paper organizes research on generalist game AI into Dataset, Model, Harness, and Benchmark pillars and charts a five-level progression from single-game mastery to agents that create and live inside game multiverses.
citing papers explorer
-
SpecBlock: Block-Iterative Speculative Decoding with Dynamic Tree Drafting
SpecBlock achieves 8-13% higher mean speedup than EAGLE-3 at 44-52% drafting cost via block-iterative drafting with hidden-state inheritance, dynamic rank-head branching, valid-prefix masking, and optional cost-aware bandit adaptation.
-
DMax: Aggressive Parallel Decoding for dLLMs
DMax uses On-Policy Uniform Training and Soft Parallel Decoding to enable aggressive parallelism in dLLMs, raising TPF on GSM8K from 2.04 to 5.47 and on MBPP from 2.71 to 5.86 while preserving accuracy.
-
Performance-Driven Policy Optimization for Speculative Decoding with Adaptive Windowing
PPOW uses window-level RL with cost-aware speedup and proximity rewards plus adaptive divergence-aware windowing to reach 6.29-6.52 acceptance lengths and 3.39-4.36x speedups in speculative decoding.
-
Dynamic Execution Commitment of Vision-Language-Action Models
A3 adaptively sets a VLA model's execution horizon to the longest action prefix the model re-confirms under conditional and sequential re-decoding, matching or improving task success without manual tuning.
-
Attention Drift: What Autoregressive Speculative Decoding Models Learn
Drafter models in speculative decoding suffer progressive attention drift caused by monotonically growing hidden-state magnitudes along the residual path; post-norm plus per-state RMSNorm reduces this drift and improves acceptance length up to 2x on perturbed templates and 1.18x on long-context data
-
Speculative Verification: Exploiting Information Gain to Refine Speculative Decoding
Speculative Verification adds a companion model that estimates draft-target alignment via information gain to dynamically set verification length, delivering up to 2x speedup over standard speculative decoding across tested models and batch sizes.
-
SpecHop: Continuous Speculation for Accelerating Multi-Hop Retrieval Agents
SpecHop accelerates multi-hop LLM tool use via continuous multi-threaded speculation with asynchronous verification, approaching oracle latency gains and reducing latency up to 40% on retrieval tasks.
-
Latent Action Reparameterization for Efficient Agent Inference
LAR learns a compact latent action space from trajectories that shortens the effective decision horizon for LLM agents, reducing token count and inference time while preserving task success.
-
Lever: Speculative LLM Inference on Smartphones
Lever optimizes the drafting, verification, and execution stages of speculative decoding for flash-backed LLM inference on smartphones, reporting 2.93x average latency reduction over baseline flash-offloaded inference.
-
Towards Generalist Game Players: An Investigation of Foundation Models in the Game Multiverse
The paper organizes research on generalist game AI into Dataset, Model, Harness, and Benchmark pillars and charts a five-level progression from single-game mastery to agents that create and live inside game multiverses.