cs.CLNov 18, 2026

NarrowSpec: Draft Hidden Width Need Not Match the Target in Speculative Decoding

Authors: Alagappan Valliappan, Ankit More, Dheevatsa Mudigere, Jatin Mitra

Organizations: NVIDIA

Abstract

Speculative decoding is widely used to speed up generation in large language models (LLMs), with a small draft model proposing tokens that one forward pass of the target model verifies. Draft models that read the target's hidden states, such as DFlash and EAGLE-3, typically inherit the target's hidden width. We show that the draft's whole residual stream can run at half or a quarter of that width while largely preserving acceptance. For DFlash, which borrows the target's vocabulary projection, the trained output adapter folds into it, giving one narrower projection with logits unchanged up to floating-point rounding. We test two drafter families, DFlash and EAGLE-3, on two targets, Qwen3-8B (NVFP4) and GPT-OSS-20B (MXFP4), in SGLang on a B200 and a DGX Spark (GB10). At batch size 1 and temperature 0, narrowing the draft to half or a quarter width speeds up end-to-end decoding by 1.06-1.12x on B200 and 1.13-1.44x on GB10. The speedup holds at larger batch sizes. From batch 32 to the largest batch each GPU fits, narrowing reaches up to 1.21x on both GPUs, and no narrow draft is slower than full width.

Figures & tables

Appendix figures & tables23 assets

Supplementary material from the paper’s appendix.

Appendix

Explore similar work

May 11, 2026

SlimSpec: Low-Rank Draft LM-Head for Accelerated Speculative Decoding

Speculative decoding speeds up autoregressive generation in Large Language Models (LLMs) through a two-step procedure, where a lightweight draft model proposes tokens which the target model then verifies in a single forward pass. Although the drafter network is small in modern architectures, its LM-head still performs projection to a large vocabulary, becoming one of the major computational bottlenecks. In prior work this issue has been predominantly addressed via static or dynamic vocabulary truncation. Yet mitigating the bottleneck, these methods bring in extra complexity, such as special vocabulary curation, sophisticated inference-time logic or modifications of the training setup. In this paper, we propose SlimSpec, a low-rank parameterization of the drafter's LM-head that compresses the inner representation rather than the output, preserving full vocabulary support. We evaluate our method with EAGLE-3 drafter across three target models and diverse benchmarks in both latency- and throughput-bound inference regimes. SlimSpec achieves 4-5×4\text{-}5\times acceleration over the standard LM-head architecture while maintaining a competitive acceptance length, surpassing existing methods by up to 8-9%8\text{-}9\% of the end-to-end speedup. Our method requires minimal adjustments of training and inference pipelines. Combined with the aforementioned speedup improvements, it makes SlimSpec a strong alternative across wide variety of draft LM-head architectures.
Jul 19, 2026

Lossless but Not Free: An Empirical Anatomy of Speculative Decoding on Consumer Hardware

Single-stream autoregressive decoding of large language models is bound by memory bandwidth: each generated token requires one full forward pass through the target model, and successive passes cannot be parallelized. Speculative decoding restructures this computation: a small draft model proposes KK tokens autoregressively, the target model scores all of them in one batched pass, and a rejection-sampling rule provably preserves the target model's output distribution. We present a from-scratch, device-agnostic (CUDA/MPS/CPU) implementation and an empirical study across five draft/target backend configurations on a consumer Apple-silicon laptop. Distribution equivalence is verified at three levels, culminating in a two-sample test over roughly 9,200 real-model tokens per method (χ2=162.5χ^2 = 162.5, dof =200= 200, p=0.976p = 0.976) and exact greedy-sequence agreement. The best configuration reaches a measured 1.61×1.61\times wall-clock speedup at K=6K=6, on an acceptance profile declining from 69.7% at K=1K=1 to 37.8% at the optimum, while three of five configurations decelerate, either because the draft fails to out-speed a small target or because the quantized Metal backend executes "parallel" verification serially, an effect we isolate and quantify. The failures are as instructive as the successes: speculative decoding pays off only when verification is genuinely batch-parallel and the draft/target latency gap is real.
Nov 3, 2026

Give the Drafter a Second Chance to Guess: Parallel Reconsideration for Speculative Decoding

Speculative decoding accelerates large language model inference with a small drafter, whose per-guess acceptance rate determines the achievable speedup. Feature-level drafters share a design choice: every token, easy or hard, is drafted under a single forward pass. Many of the tokens the target rejects are ones the drafter is itself unsure about yet the target answers decisively. To address this, we introduce Parallel Reconsideration: guided by its own uncertainty, the drafter reconsiders its hesitant guesses with extra passes that run in parallel with the one-pass draft, adding negligible latency. A co-designed training recipe and serving-time drafting algorithm append these reconsidered guesses to the original draft tree as refined candidates. Across models and benchmarks, on average, Parallel Reconsideration improves over the baselines by 5.3% to 10.2% on acceptance length and by 3.5% to 9.2% on speedup.