Pith. sign in

REVIEW 6 cited by

Recursive Speculative Decoding: Accelerating LLM Inference via Sampling Without Replacement

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2402.14160 v2 pith:OVOA424S submitted 2024-02-21 cs.LG cs.AI

Recursive Speculative Decoding: Accelerating LLM Inference via Sampling Without Replacement

classification cs.LG cs.AI
keywords treecomputationaldecodingfixedreplacementspeculativewithoutdraft
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
Share X Bluesky LinkedIn Reddit HN
abstract

Speculative decoding is an inference-acceleration method for large language models (LLMs) where a small language model generates a draft-token sequence which is further verified by the target LLM in parallel. Recent works have advanced this method by establishing a draft-token tree, achieving superior performance over a single-sequence speculative decoding. However, those works independently generate tokens at each level of the tree, not leveraging the tree's entire diversifiability. Besides, their empirical superiority has been shown for fixed length of sequences, implicitly granting more computational resource to LLM for the tree-based methods. None of the existing works has conducted empirical studies with fixed target computational budgets despite its importance to resource-bounded devices. We present Recursive Speculative Decoding (RSD), a novel tree-based method that samples draft tokens without replacement and maximizes the diversity of the tree. During RSD's drafting, the tree is built by either Gumbel-Top-$k$ trick that draws tokens without replacement in parallel or Stochastic Beam Search that samples sequences without replacement while early-truncating unlikely draft sequences and reducing the computational cost of LLM. We empirically evaluate RSD with Llama 2 and OPT models, showing that RSD outperforms the baseline methods, consistently for fixed draft sequence length and in most cases for fixed computational budgets at LLM.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 6 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Cost-Aware Diffusion Draft Trees for Speculative Decoding

    cs.CL 2026-06 unverdicted novelty 7.0

    CaDDTree jointly selects tree structure and budget to maximize expected tokens per unit time in speculative decoding, proving unimodality under convex verification cost and matching oracle DDTree performance on Qwen models.

  2. UniVer: A Unified Perspective for Multi-step and Multi-draft Speculative Decoding

    cs.CL 2026-05 unverdicted novelty 7.0

    UniVer frames tree-based speculative decoding as conditional optimal transport, proving it is lossless with optimal acceptance rates and delivering 4.2-8.5% longer accepted sequences than standard rejection sampling.

  3. Cactus: Accelerating Auto-Regressive Decoding with Constrained Acceptance Speculative Sampling

    cs.LG 2026-04 unverdicted novelty 7.0

    Cactus uses constrained optimization to guarantee bounded divergence from the verifier LLM distribution during speculative sampling, raising acceptance rates without the distortion seen in typical acceptance sampling.

  4. Structuring The Future: Diffusion LLM Speculative Decoding via Calibrated Draft Graphs

    cs.LG 2025-09 conditional novelty 6.0

    Spiffy speeds up diffusion LLM inference up to about 3x (and up to 7.9x with parallel decoding) by verifying multiple candidate unmasked states in one batched model call, while preserving greedy output.

  5. ConFu: Contemplate the Future for Better Speculative Sampling

    cs.CL 2026-03 unverdicted novelty 5.0

    ConFu boosts speculative decoding acceptance rates 8-20% over EAGLE-3 by letting draft models use contemplate tokens and MoE to anticipate future generation direction.

  6. Efficient Speculative Decoding for Llama at Scale: Challenges and Solutions

    cs.CL 2025-08 conditional novelty 5.0

    Meta reports EAGLE-based speculative decoding optimizations for Llama models, achieving state-of-the-art latency (about 4 ms/token for Llama4 Maverick) and 1.4-2.0x speedups for large batches.