The paper investigates speculative decoding, a technique that speeds up language model inference by generating draft tokens and verifying them. It shows that when verification is relaxed (lossy), the resulting completions become longer—especially due to extended reasoning traces—despite accepting more tokens per round. The study evaluates various relaxed verification rules across multiple model pairs, finding that higher acceptance rates often lead to longer outputs, lower target-model probability tokens, and reduced accuracy, indicating that token acceptance alone is an unreliable metric for decoding speed.
arXiv:2608.30427v1 Announce Type: cross Abstract: Speculative decoding speeds up generation with an efficient draft model (drafter) that proposes tokens for a target model to verify in one pass, pres...
The paper shows that long‑context drafting can be performed with short‑context drafters that are frozen and run beyond their training length, using two inference‑time mechanisms to keep their cost bounded. By bounding the attention cache and re‑basing rotary position indices, the authors demonstrate that the speedup achieved with a short‑context drafter persists up to two million tokens across multiple target‑drafter pairs and corpora. They prove that when a common index shift leaves the drafter’s computation unchanged, the acceptance gap can be tightly bounded, ensuring consistent performance.
arXiv:2609.14717v1 Announce Type: cross Abstract: Speculative decoding accelerates LLM inference by verifying multiple drafted tokens in parallel, allowing a single target forward pass to accept seve...
The paper introduces APEX, a learned controller that improves speculative decoding for large language models by dynamically selecting the best speculation strategy and adjusting draft depth during generation. APEX uses request‑level routing among EAGLE‑3, n‑gram, and draft‑model speculation, and block‑level depth adaptation based on causal decoding signals and verifier feedback. Integrated into vLLM and tested with Qwen3‑8B, APEX achieves up to 5.24× speedup over autoregressive decoding while reducing wasted tokens by 41% compared to fixed‑depth n‑gram speculation.
arXiv:2610.07822v1 Announce Type: new Abstract: Speculative decoding accelerates autoregressive generation by using a lightweight draft model to propose multiple tokens that are verified by a target...
arXiv:2608. 03447v1 Announce Type: cross Abstract: Speculative decoding accelerates autoregressive generation by verifying a draft block with a target model in parallel.
The paper introduces AdaptiveSpec, a training‑free speculative decoding method that simultaneously adapts the per‑step verification rule and the draft‑tree shape using signals generated during decoding. It replaces the fixed token‑match rule with a margin‑based threshold and adjusts tree depth, width, and node count based on draft confidence and recent acceptance history, allowing the total draft count to vary. Experiments on SGLang show up to 56% throughput gains over EAGLE‑3 while maintaining 93% of lossless task accuracy on GSM8K, MATH‑500, and HumanEval across three models.
arXiv:2608. 03401v1 Announce Type: cross Abstract: Large language models often reason at length before answering, increasing cost and latency.
arXiv:2609.24150v1 Announce Type: new Abstract: Speculative decoding accelerates large language model (LLM) inference by using a lightweight draft model to generate multiple candidate tokens that are...
arXiv:2606. 01019v1 Announce Type: cross Abstract: Large Language Model (LLM) generation remains expensive because autoregressive decoding calls the model once for each new token.
Verification-Aware Training (VAT) is a plug‑in framework that improves speculative decoding for large language models by simulating verification during training and using the resulting accept/reject patterns as supervision. VAT adds a lightweight binary verification head to predict whether each draft token will survive sequential verification, and replaces the fixed per‑position weighting with a verification‑adaptive schedule that keeps full weight up to the first rejection point. When applied to EAGLE‑3 and DFlash on Qwen3‑4B, Qwen3‑8B, and LLaMA‑3.1‑8B, VAT increases average acceptance length by up to 11.4% and wall‑clock speedup by up to 8.7%, yielding consistent gains across math, code, and chat benchmarks.
The paper explores a new lossy strategy for speculative decoding called verifier skipping, which allows certain draft prefixes to be committed without invoking the verifier. It investigates which confidence signals—raw token confidence, learned marginal survival, or learned conditional survival—best schedule these skips, finding that raw confidence often performs best despite higher positionwise AUROC for learned signals. Experiments on HumanEval with DiffuCoder-7B-Instruct and Qwen3-32B show that all three signals can reduce verifier calls by 9.6% to 13.5% while maintaining the same pass@1 as Strict SDD.