Daily Papers

Speculative decoding accelerates autoregressive generation by having a cheap draft propose tokens that a target verifies in parallel. Frontier models increasingly ship a built-in Multi-Token-Prediction (MTP/NEXTN) draft head under the assumption that the draft is negligibly cheap. At million-token context this breaks: an MTP draft head typically runs full attention over the entire KV cache at every draft step, so its read grows linearly with context and comes to dominate the draft cost -- precisely where speculation is most valuable. The effect compounds with draft length (a deep native draft can turn net-negative, slower than no speculation) and sharpens under hybrid/linear-attention targets, where cheaper verification leaves the draft's full-attention read exposed. We apply a StreamingLLM-style sliding window plus attention sink to the draft's attention only (Windowed-MTP), leaving full-attention verification intact. It is training-free, drop-in, and lossless by construction: the full-attention target still decides every accepted token, so windowing changes only which tokens are proposed, never which are accepted. It bounds the draft's KV working set to a constant, dropping ~99% of KV entries at 1M. Across three architecture families (Qwen GDN-MoE 35B/122B and a Mamba2-hybrid NoPE 120B) at 1M context on a single GPU in SGLang, windowing cuts the per-decode-step cost over the shipping native MTP draft by +28% to +44%, an input-invariant margin that widens with context. Since per-token latency is this cost divided by acceptance length, at matched acceptance end-to-end decode latency improves by the same amount, and more where windowing also lifts acceptance, while preserving the target's verified output distribution. Finally, the unread draft KV -- 7.7-11% of total KV at 1M -- is reclaimed via a compact ring buffer at no acceptance or quality cost.

Sustainability disclosure standards (e.g., GRI, SASB, TCFD, IFRS S2) are comprehensive yet lengthy, terminology-dense, and highly cross-referential, hindering structured analysis and downstream use. We present SSKG Hub (Sustainability Standards Knowledge Graph Hub), a research prototype and interactive web platform that transforms standards into auditable knowledge graphs (KGs) through an LLM-centered, expert-guided pipeline. The system integrates automatic standard identification, configurable chunking, standard-specific prompting, robust triple parsing, and provenance-aware Neo4j storage with fine-grained audit metadata. LLM extraction produces a provenance-linked Draft KG, which is reviewed, curated, and formally promoted to a Certified KG through meta-expert adjudication. A role-based governance framework covering read-only guest access, expert review and CRUD operations, meta-expert certification, and administrative oversight ensures traceability and accountability across draft and certified states. Beyond graph exploration and triple-level evidence tracing, SSKG Hub supports cross-KG fusion, KG-driven tasks, and dedicated modules for insights and curated resources. We validate the platform through a comprehensive expert-led KG review case study that demonstrates end-to-end curation and quality assurance. The web application is publicly available at www.sskg-hub.com.

We present Direct Reward Fine-Tuning (DRaFT), a simple and effective method for fine-tuning diffusion models to maximize differentiable reward functions, such as scores from human preference models. We first show that it is possible to backpropagate the reward function gradient through the full sampling procedure, and that doing so achieves strong performance on a variety of rewards, outperforming reinforcement learning-based approaches. We then propose more efficient variants of DRaFT: DRaFT-K, which truncates backpropagation to only the last K steps of sampling, and DRaFT-LV, which obtains lower-variance gradient estimates for the case when K=1. We show that our methods work well for a variety of reward functions and can be used to substantially improve the aesthetic quality of images generated by Stable Diffusion 1.4. Finally, we draw connections between our approach and prior work, providing a unifying perspective on the design space of gradient-based fine-tuning algorithms.

Speculative Decoding (SD) has become an important technique in accelerating the inference speed of large language models. Conventional SD methods employ a fixed draft length, which ignores the token generation difficulty across tasks. Consequently, in this paper, we address such an issue and introduce SVIP - a difficulty-aware dynamic draft length policy for speculative decoding systems. Based on a theoretical lower bound of draft token acceptance rate and its inference-time approximation, SVIP adaptively determines the lengths of draft sequences based on the entropy of each draft token distribution. Experimental results on mainstream SD benchmarks and frameworks demonstrate the superior performance of SVIP, achieving up to 20\% walltime speedup on SpecBench over baseline SD methods and 60\% speedup on MT-Bench for long-form generation of up to 8K tokens. Moreover, SVIP is totally training-free and compatible with any existing SD methods that generate draft tokens autoregressively. Experimental results also show that SVIP yields consistent walltime improvement on top of GliDe & CaPE and EAGLE-2.

Speculative decoding accelerates LLM inference, but SOTA hidden-state-based drafters suffer from long-range decay: draft accuracy degrades as the speculative step increases. Existing work attributes this decay to train-inference mismatch and proposes test-time training (TTT) as a remedy, yet we observe that long-range decay persists even in TTT-trained drafters. We revisit long-range decay from the perspective of context information preservation. In hidden-state reuse, we argue the target hidden state acts as a biased context compression: it aggregates historical token information according to the attention query at the current position, yielding a compact representation optimized for immediate next-token prediction. This compression can suppress information less relevant to the current query but important for later speculative steps. In contrast, the target model's KV cache serves as an explicit context, retaining the complete set of token-wise KV representations. We therefore posit the KV-Reuse Hypothesis: allowing the draft model to reuse the target KV cache can provide richer signals for long-horizon drafting. To test this hypothesis, we introduce KVShot, a diagnostic framework that compares three reuse paradigms: hidden-only, KV-only, and hybrid. Extensive evaluations on Qwen3-8B show that KV-Reuse improves long-range acceptance, although end-to-end speedups remain marginal under current training pipelines. Our analysis identifies two key structural bottlenecks: shallow drafters struggle to estimate target queries accurately, and draft-side KV projections receive sparse gradient signals. These findings suggest that realizing the full potential of KV-aware decoding requires moving beyond TTT toward block-wise training paradigms. By exposing these bottlenecks, KVShot provides a foundational diagnostic testbed and a clear roadmap for designing next-generation inference architectures.

Speculative decoding is an effective and lossless method for Large Language Model (LLM) inference acceleration. It employs a smaller model to generate a draft token sequence, which is then verified by the original base model. In multi-GPU systems, inference latency can be further reduced through tensor parallelism (TP), while the optimal TP size of the draft model is typically smaller than that of the base model, leading to GPU idling during the drafting stage. To solve this problem, we propose EasySpec, a layer-parallel speculation strategy that optimizes the efficiency of multi-GPU utilization.EasySpec breaks the sequential execution order of layers in the drafting model, enabling multi-layer parallelization across devices, albeit with some induced approximation errors. After each drafting-and-verification iteration, the draft model's key-value (KV) cache is calibrated in a single forward pass, preventing long-term error accumulation at minimal additional latency. We evaluated EasySpec on several mainstream open-source LLMs, using smaller versions of models from the same series as drafters. The results demonstrate that EasySpec can achieve a peak speedup of 4.17x compared to vanilla decoding, while preserving the original distribution of the base LLMs. Specifically, the drafting stage can be accelerated by up to 1.62x with a maximum accuracy drop of only 7%, requiring no training or fine-tuning on the draft models.

Diffusion transformer-based video generation models (DiTs) have recently attracted widespread attention for their excellent generation quality. However, their computational cost remains a major bottleneck-attention alone accounts for over 80% of total latency, and generating just 8 seconds of 720p video takes tens of minutes-posing serious challenges to practical application and scalability. To address this, we propose the DraftAttention, a training-free framework for the acceleration of video diffusion transformers with dynamic sparse attention on GPUs. We apply down-sampling to each feature map across frames in the compressed latent space, enabling a higher-level receptive field over the latent composed of hundreds of thousands of tokens. The low-resolution draft attention map, derived from draft query and key, exposes redundancy both spatially within each feature map and temporally across frames. We reorder the query, key, and value based on the draft attention map to guide the sparse attention computation in full resolution, and subsequently restore their original order after the attention computation. This reordering enables structured sparsity that aligns with hardware-optimized execution. Our theoretical analysis demonstrates that the low-resolution draft attention closely approximates the full attention, providing reliable guidance for constructing accurate sparse attention. Experimental results show that our method outperforms existing sparse attention approaches in video generation quality and achieves up to 1.75x end-to-end speedup on GPUs. Code: https://github.com/shawnricecake/draft-attention

Large Language Models (LLMs) frequently suffer from inefficiencies, largely attributable to the discord between the requirements of auto-regressive decoding and the architecture of contemporary GPUs. Recently, regressive lightweight speculative decoding has garnered attention for its notable efficiency improvements in text generation tasks. This approach utilizes a lightweight regressive draft model, like a Recurrent Neural Network (RNN) or a single transformer decoder layer, leveraging sequential information to iteratively predict potential tokens. Specifically, RNN draft models are computationally economical but tend to deliver lower accuracy, while attention decoder layer models exhibit the opposite traits. This paper presents Clover-2, an advanced iteration of Clover, an RNN-based draft model designed to achieve comparable accuracy to that of attention decoder layer models while maintaining minimal computational overhead. Clover-2 enhances the model architecture and incorporates knowledge distillation to increase Clover's accuracy and improve overall efficiency. We conducted experiments using the open-source Vicuna 7B and LLaMA3-Instruct 8B models. The results demonstrate that Clover-2 surpasses existing methods across various model architectures, showcasing its efficacy and robustness.

Large Language Models (LLMs) exhibit high inference latency due to their autoregressive decoding nature. While the draft head in speculative decoding mitigates this issue, its full potential remains unexplored. In this paper, we introduce KOALA (K-layer Optimized Adversarial Learning Architecture), an orthogonal approach to the draft head. By transforming the conventional single-layer draft head into a multi-layer architecture and incorporating adversarial learning into the traditional supervised training, KOALA significantly improves the accuracy of the draft head in predicting subsequent tokens, thus more closely mirroring the functionality of LLMs. Although this improvement comes at the cost of slightly increased drafting overhead, KOALA substantially unlocks the draft head's potential, greatly enhancing speculative decoding. We conducted comprehensive evaluations of KOALA, including both autoregressive and non-autoregressive draft heads across various tasks, demonstrating a latency speedup ratio improvement of 0.24x-0.41x, which is 10.57%-14.09% faster than the original draft heads.

Architectural Knowledge Management (AKM) is crucial for software development but remains challenging due to the lack of standardization and high manual effort. Architecture Decision Records (ADRs) provide a structured approach to capture Architecture Design Decisions (ADDs), but their adoption is limited due to the manual effort involved and insufficient tool support. Our previous work has shown that Large Language Models (LLMs) can assist in generating ADDs. However, simply prompting the LLM does not produce quality ADDs. Moreover, using third-party LLMs raises privacy concerns, while self-hosting them poses resource challenges. To this end, we experimented with different approaches like few-shot, retrieval-augmented generation (RAG) and fine-tuning to enhance LLM's ability to generate ADDs. Our results show that both techniques improve effectiveness. Building on this, we propose Domain Specific Retreival Augumented Few Shot Fine Tuninng, DRAFT, which combines the strengths of all these three approaches for more effective ADD generation. DRAFT operates in two phases: an offline phase that fine-tunes an LLM on generating ADDs augmented with retrieved examples and an online phase that generates ADDs by leveraging retrieved ADRs and the fine-tuned model. We evaluated DRAFT against existing approaches on a dataset of 4,911 ADRs and various LLMs and analyzed them using automated metrics and human evaluations. Results show DRAFT outperforms all other approaches in effectiveness while maintaining efficiency. Our findings indicate that DRAFT can aid architects in drafting ADDs while addressing privacy and resource constraints.

Optimizing inference for long-context Large Language Models (LLMs) is increasingly important due to the quadratic compute and linear memory complexity of Transformers. Existing approximation methods, such as key-value (KV) cache dropping, sparse attention, and prompt compression, typically rely on rough predictions of token or KV pair importance. We propose a novel framework for approximate LLM inference that leverages small draft models to more accurately predict the importance of tokens and KV pairs. Specifically, we introduce two instantiations of our proposed framework: (i) SpecKV, which leverages a draft output to accurately assess the importance of each KV pair for more effective KV cache dropping, and (ii) SpecPC, which uses the draft model's attention activations to identify and discard unimportant prompt tokens. To the best of our knowledge, this is the first work to use draft models for approximate LLM inference acceleration, extending their utility beyond traditional lossless speculative decoding. We motivate our methods with theoretical and empirical analyses, and show a strong correlation between the attention patterns of draft and target models. Extensive experiments on long-context benchmarks show that our methods consistently achieve higher accuracy than existing baselines, while preserving the same improvements in memory usage, latency, and throughput. Our code is available at https://github.com/furiosa-ai/draft-based-approx-llm.

We consider multi-draft speculative sampling, where the proposal sequences are sampled independently from different draft models. At each step, a token-level draft selection scheme takes a list of valid tokens as input and produces an output token whose distribution matches that of the target model. Previous works have demonstrated that the optimal scheme (which maximizes the probability of accepting one of the input tokens) can be cast as a solution to a linear program. In this work we show that the optimal scheme can be decomposed into a two-step solution: in the first step an importance sampling (IS) type scheme is used to select one intermediate token; in the second step (single-draft) speculative sampling is applied to generate the output token. For the case of two identical draft models we further 1) establish a necessary and sufficient condition on the distributions of the target and draft models for the acceptance probability to equal one and 2) provide an explicit expression for the optimal acceptance probability. Our theoretical analysis also motives a new class of token-level selection scheme based on weighted importance sampling. Our experimental results demonstrate consistent improvements in the achievable block efficiency and token rates over baseline schemes in a number of scenarios.

Large language models (LLMs) are increasingly used as co-authors in collaborative writing, where users begin with rough drafts and rely on LLMs to complete, revise, and refine their content. However, this capability poses a serious safety risk: malicious users could jailbreak the models-filling incomplete drafts with dangerous content-to force them into generating harmful outputs. In this paper, we identify the vulnerability of current LLMs to such draft-based co-authoring jailbreak attacks and introduce HarDBench, a systematic benchmark designed to evaluate the robustness of LLMs against this emerging threat. HarDBench spans a range of high-risk domains-including Explosives, Drugs, Weapons, and Cyberattacks-and features prompts with realistic structure and domain-specific cues to assess the model susceptibility to harmful completions. To mitigate this risk, we introduce a safety-utility balanced alignment approach based on preference optimization, training models to refuse harmful completions while remaining helpful on benign drafts. Experimental results show that existing LLMs are highly vulnerable in co-authoring contexts and our alignment method significantly reduces harmful outputs without degrading performance on co-authoring capabilities. This presents a new paradigm for evaluating and aligning LLMs in human-LLM collaborative writing settings. Our new benchmark and dataset are available on our project page at https://github.com/untae0122/HarDBench

We present a PaperRobot who performs as an automatic research assistant by (1) conducting deep understanding of a large collection of human-written papers in a target domain and constructing comprehensive background knowledge graphs (KGs); (2) creating new ideas by predicting links from the background KGs, by combining graph attention and contextual text attention; (3) incrementally writing some key elements of a new paper based on memory-attention networks: from the input title along with predicted related entities to generate a paper abstract, from the abstract to generate conclusion and future work, and finally from future work to generate a title for a follow-on paper. Turing Tests, where a biomedical domain expert is asked to compare a system output and a human-authored string, show PaperRobot generated abstracts, conclusion and future work sections, and new titles are chosen over human-written ones up to 30%, 24% and 12% of the time, respectively.

Although autoregressive models have achieved promising results on image generation, their unidirectional generation process prevents the resultant images from fully reflecting global contexts. To address the issue, we propose an effective image generation framework of Draft-and-Revise with Contextual RQ-transformer to consider global contexts during the generation process. As a generalized VQ-VAE, RQ-VAE first represents a high-resolution image as a sequence of discrete code stacks. After code stacks in the sequence are randomly masked, Contextual RQ-Transformer is trained to infill the masked code stacks based on the unmasked contexts of the image. Then, Contextual RQ-Transformer uses our two-phase decoding, Draft-and-Revise, and generates an image, while exploiting the global contexts of the image during the generation process. Specifically. in the draft phase, our model first focuses on generating diverse images despite rather low quality. Then, in the revise phase, the model iteratively improves the quality of images, while preserving the global contexts of generated images. In experiments, our method achieves state-of-the-art results on conditional image generation. We also validate that the Draft-and-Revise decoding can achieve high performance by effectively controlling the quality-diversity trade-off in image generation.

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 4-5times acceleration over the standard LM-head architecture while maintaining a competitive acceptance length, surpassing existing methods by up to 8-9% 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.

Speculative decoding (SD) accelerates large language model inference by leveraging a draft-then-verify paradigm. To maximize the acceptance rate, recent methods construct expansive draft trees, which unfortunately incur severe VRAM bandwidth and computational overheads that bottleneck end-to-end speedups. While dynamic-depth pruning can reduce this latency by removing marginal branches, it also discards potentially valid candidates, preventing the acceptance rate from reaching the upper bound of dense trees. In this paper, we identify a critical opportunity in resource allocation: the transition from dense to pruned drafting frees up significant computational budget. To break this Pareto tradeoff, we introduce Graft, a compensation framework that couples pruning and retrieval as mutually reinforcing operations. Pruning supplies sufficient budget for retrieval, while retrieval compensates for pruning-induced coverage loss and recovers accepted length. By employing a sequential `prune-then-graft' mechanism, Graft attaches highly predictive retrieved tokens into positions opened by pruning, filling the topological gaps with near-zero overhead. Graft is entirely training-free and lossless. Comprehensive evaluations show that Graft establishes a new Pareto frontier across practical deployment settings, including short-context generation, long-context generation, and large-scale models. On short-context benchmarks, it achieves up to 5.41times speedup and improves average speedup over EAGLE-3 by up to 21.8% on the large-scale Qwen3-235B. We also provide a preliminary exploration of applying Graft to the DFlash-style block drafting paradigm, offering initial evidence and insights for extending grafting beyond autoregressive draft trees.

We present a novel inference scheme, self-speculative decoding, for accelerating Large Language Models (LLMs) without the need for an auxiliary model. This approach is characterized by a two-stage process: drafting and verification. The drafting stage generates draft tokens at a slightly lower quality but more quickly, which is achieved by selectively skipping certain intermediate layers during drafting Subsequently, the verification stage employs the original LLM to validate those draft output tokens in one forward pass. This process ensures the final output remains identical to that produced by the unaltered LLM, thereby maintaining output quality. The proposed method requires no additional neural network training and no extra memory footprint, making it a plug-and-play and cost-effective solution for inference acceleration. Benchmarks with LLaMA-2 and its fine-tuned models demonstrated a speedup up to 1.73times.

This paper presents a personalized character recommendation system for Multiplayer Online Battle Arena (MOBA) games which are considered as one of the most popular online video game genres around the world. When playing MOBA games, players go through a draft stage, where they alternately select a virtual character to play. When drafting, players select characters by not only considering their character preferences, but also the synergy and competence of their team's character combination. However, the complexity of drafting induces difficulties for beginners to choose the appropriate characters based on the characters of their team while considering their own champion preferences. To alleviate this problem, we propose DraftRec, a novel hierarchical model which recommends characters by considering each player's champion preferences and the interaction between the players. DraftRec consists of two networks: the player network and the match network. The player network captures the individual player's champion preference, and the match network integrates the complex relationship between the players and their respective champions. We train and evaluate our model from a manually collected 280,000 matches of League of Legends and a publicly available 50,000 matches of Dota2. Empirically, our method achieved state-of-the-art performance in character recommendation and match outcome prediction task. Furthermore, a comprehensive user survey confirms that DraftRec provides convincing and satisfying recommendations. Our code and dataset are available at https://github.com/dojeon-ai/DraftRec.

To combat the memory bandwidth-bound nature of autoregressive LLM inference, previous research has proposed the speculative decoding framework. To perform speculative decoding, a small draft model proposes candidate continuations of the input sequence, that are then verified in parallel by the base model. One way to specify the draft model, as used in the recent Medusa decoding framework, is as a collection of light-weight heads, called draft heads, that operate on the base model's hidden states. To date, all existing draft heads have been sequentially independent, meaning that they speculate tokens in the candidate continuation independently of any preceding tokens in the candidate continuation. In this work, we propose Hydra heads, a sequentially dependent, drop-in replacement for standard draft heads that significantly improves speculation accuracy. Decoding with Hydra heads improves throughput compared to Medusa decoding with standard draft heads. We further explore the design space of Hydra head training objectives and architectures, and propose a carefully-tuned Hydra head recipe, which we call Hydra++, that improves decoding throughput by 1.31x and 2.71x compared to Medusa decoding and autoregressive decoding, respectively. Overall, Hydra heads are a simple intervention on standard draft heads that significantly improve the end-to-end speed of draft head based speculative decoding.

We propose TETRIS, a novel method that optimizes the total throughput of batch speculative decoding in multi-request settings. Unlike existing methods that optimize for a single request or a group of requests as a whole, TETRIS actively selects the most promising draft tokens (for every request in a batch) to be accepted when verified in parallel, resulting in fewer rejected tokens and hence less wasted computing resources. Such an effective resource utilization to achieve fast inference in large language models (LLMs) is especially important to service providers with limited inference capacity. Compared to baseline speculative decoding, TETRIS yields a consistently higher acceptance rate and more effective utilization of the limited inference capacity. We show theoretically and empirically that TETRIS outperforms baseline speculative decoding and existing methods that dynamically select draft tokens, leading to a more efficient batch inference in LLMs.

Speculative Decoding (SD) is a popular lossless technique for accelerating the inference of Large Language Models (LLMs). We show that the decoding speed of SD frameworks with static draft structures can be significantly improved by incorporating context-aware adaptive draft structures. However, current studies on adaptive draft structures are limited by their performance, modeling approaches, and applicability. In this paper, we introduce AdaEAGLE, the first SD framework that explicitly models adaptive draft structures. AdaEAGLE leverages the Lightweight Draft Length Predictor (LDLP) module to explicitly predict the optimal number of draft tokens during inference to guide the draft model. It achieves comparable speedup results without manual thresholds and allows for deeper, more specialized optimizations. Moreover, together with threshold-based strategies, AdaEAGLE achieves a 1.62times speedup over the vanilla AR decoding and outperforms fixed-length SotA baseline while maintaining output quality.

Speculative decoding (SD), where an extra draft model is employed to provide multiple draft tokens first and then the original target model verifies these tokens in parallel, has shown great power for LLM inference acceleration. However, existing SD methods suffer from the mutual waiting problem, i.e., the target model gets stuck when the draft model is guessing tokens, and vice versa. This problem is directly incurred by the asynchronous execution of the draft model and the target model, and is exacerbated due to the fixed draft length in speculative decoding. To address these challenges, we propose a conceptually simple, flexible, and general framework to boost speculative decoding, namely Parallel spEculative decoding with Adaptive dRaft Length (PEARL). Specifically, PEARL proposes pre-verify to verify the first draft token in advance during the drafting phase, and post-verify to generate more draft tokens during the verification phase. PEARL parallels the drafting phase and the verification phase via applying the two strategies, and achieves adaptive draft length for different scenarios, which effectively alleviates the mutual waiting problem. Moreover, we theoretically demonstrate that the mean accepted tokens of PEARL is more than existing draft-then-verify works. Experiments on various text generation benchmarks demonstrate the effectiveness of our \name, leading to a superior speedup performance up to 3.79times and 1.52times, compared to auto-regressive decoding and vanilla speculative decoding, respectively.

Large language models (LLMs) are increasingly used to generate executable outputs, JSON objects, and API calls, where a single syntax error can make the output unusable. Constrained decoding enforces validity token-by-token via masking and renormalization, but it can distort generation when the model assigns low probability mass to valid continuations, pushing decoding toward locally valid yet semantically incorrect trajectories. We propose Draft-Conditioned Constrained Decoding (DCCD), a simple two-step, training-free inference procedure that decouples semantic planning from structural enforcement: an unconstrained draft is generated first, and constrained decoding is then applied, conditioned on this draft, to guarantee validity. We analyze DCCD through a KL-projection view, showing that draft conditioning increases feasible mass and reduces the cumulative "projection tax" induced by hard constraints, with an optional best-of-K draft selection. Across structured reasoning benchmarks, DCCD improves strict structured accuracy by up to +24 percentage points over standard constrained decoding (e.g., 15.2\% to 39.0\% on GSM8K with a 1B model), and enables smaller model pairs to match or exceed much larger constrained baselines, yielding substantial gains in parameter efficiency.

Speculative decoding has emerged as a pivotal technique to accelerate LLM inference by employing a lightweight draft model to generate candidate tokens that are subsequently verified by the target model in parallel. However, while this paradigm successfully increases the arithmetic intensity of memory-bound inference, it causes significant compute inefficiency: the majority of draft tokens fail verification and are discarded, resulting in waste of computation. Motivated by the goal of recollecting this wasted computation, we propose a novel system that transforms discarded drafts into reusable tokens. Our key insight is to perform auto-regressive prediction at the hidden states level and postpone the integrating token information after the hidden states generation, so the draft hidden states are not contaminated by incorrect tokens, enabling hidden state reuse. To implement such a system, first we introduce a draft model architecture based on auto-regressive hidden states, which preserves richer semantics than token-based drafters to facilitate draft repurposing. Second, we design an efficient token information injection mechanism that leverages our specialized draft model to construct high-quality draft token trees and enables resampling tokens from verification failures. Third, we eliminate the overhead hidden in our design to further maximize hardware utilization. We conducted extensive evaluations against various baselines, demonstrating up to a 3.3x speedup against standard speculative decoding.

Speculative decoding is an effective method for accelerating inference of large language models (LLMs) by employing a small draft model to predict the output of a target model. However, when adapting speculative decoding to domain-specific target models, the acceptance rate of the generic draft model drops significantly due to domain shift. In this work, we systematically investigate knowledge distillation techniques for training domain draft models to improve their speculation accuracy. We compare white-box and black-box distillation approaches and explore their effectiveness in various data accessibility scenarios, including historical user queries, curated domain data, and synthetically generated alignment data. Our experiments across Function Calling, Biology, and Chinese domains show that offline distillation consistently outperforms online distillation by 11% to 25%, white-box distillation surpasses black-box distillation by 2% to 10%, and data scaling trends hold across domains. Additionally, we find that synthetic data can effectively align draft models and achieve 80% to 93% of the performance of training on historical user queries. These findings provide practical guidelines for training domain-specific draft models to improve speculative decoding efficiency.

Speculative Decoding (SD) accelerates large language model inference by employing a small draft model to generate predictions, which are then verified by a larger target model. The effectiveness of SD hinges on the alignment between these models, which is typically enhanced by Knowledge Distillation (KD). However, conventional KD methods aim to minimize the KL divergence between the draft and target models across all tokens, a goal that is misaligned with the true objective of SD, which is to maximize token acceptance rate. Therefore, draft models often struggle to fully assimilate the target model's knowledge due to capacity constraints, leading to suboptimal performance. To address this challenge, we propose AdaSPEC, a novel method that incorporates selective token filtering into the KD process. AdaSPEC utilizes a reference model to identify and filter out difficult-to-fit tokens, enabling the distillation of a draft model that better aligns with the target model on simpler tokens. This approach improves the overall token acceptance rate without compromising generation quality. We evaluate AdaSPEC across diverse tasks, including arithmetic reasoning, instruction-following, coding, and summarization, using model configurations of 31M/1.4B and 350M/2.7B parameters. Our results demonstrate that AdaSPEC consistently outperforms the state-of-the-art DistillSpec method, achieving higher acceptance rates across all tasks (up to 15\%). The code is publicly available at https://github.com/yuezhouhu/adaspec.

Speculative decoding has demonstrated its effectiveness in accelerating the inference of large language models while maintaining a consistent sampling distribution. However, the conventional approach of training a separate draft model to achieve a satisfactory token acceptance rate can be costly. Drawing inspiration from early exiting, we propose a novel self-speculative decoding framework Kangaroo, which uses a fixed shallow sub-network as a self-draft model, with the remaining layers serving as the larger target model. We train a lightweight and efficient adapter module on top of the sub-network to bridge the gap between the sub-network and the full model's representation ability. It is noteworthy that the inference latency of the self-draft model may no longer be negligible compared to the large model, necessitating strategies to increase the token acceptance rate while minimizing the drafting steps of the small model. To address this challenge, we introduce an additional early exiting mechanism for generating draft tokens. Specifically, we halt the small model's subsequent prediction during the drafting phase once the confidence level for the current token falls below a certain threshold. Extensive experiments on the Spec-Bench demonstrate the effectiveness of Kangaroo. Under single-sequence verification, Kangaroo achieves speedups up to 1.68times on Spec-Bench, outperforming Medusa-1 with 88.7\% fewer additional parameters (67M compared to 591M). The code for Kangaroo is available at https://github.com/Equationliu/Kangaroo.

Speculative Decoding (SD) has emerged as a widely used paradigm to accelerate the inference of large language models (LLMs) without compromising generation quality. It works by efficiently drafting multiple tokens using a compact model and then verifying them in parallel using the target LLM. Notably, Self-Speculative Decoding proposes skipping certain layers to construct the draft model, which eliminates the need for additional parameters or training. Despite its strengths, we observe in this work that drafting with layer skipping exhibits significant sensitivity to domain shifts, leading to a substantial drop in acceleration performance. To enhance the domain generalizability of this paradigm, we introduce KNN-SSD, an algorithm that leverages K-Nearest Neighbor (KNN) search to match different skipped layers with various domain inputs. We evaluated our algorithm in various models and multiple tasks, observing that its application leads to 1.3x-1.6x speedup in LLM inference.

Text generation with Large Language Models (LLMs) is known to be memory bound due to the combination of their auto-regressive nature, huge parameter counts, and limited memory bandwidths, often resulting in low token rates. Speculative decoding has been proposed as a solution for LLM inference acceleration. However, since draft models are often unavailable in the modern open-source LLM families, e.g., for Llama 2 7B, training a high-quality draft model is required to enable inference acceleration via speculative decoding. In this paper, we propose a simple draft model training framework for direct alignment to chat-capable target models. With the proposed framework, we train Llama 2 Chat Drafter 115M, a draft model for Llama 2 Chat 7B or larger, with only 1.64\% of the original size. Our training framework only consists of pretraining, distillation dataset generation, and finetuning with knowledge distillation, with no additional alignment procedure. For the finetuning step, we use instruction-response pairs generated by target model for distillation in plausible data distribution, and propose a new Total Variation Distance++ (TVD++) loss that incorporates variance reduction techniques inspired from the policy gradient method in reinforcement learning. Our empirical results show that Llama 2 Chat Drafter 115M with speculative decoding achieves up to 2.3 block efficiency and 2.4times speed-up relative to autoregressive decoding on various tasks with no further task-specific fine-tuning.

The large size of the KV cache has become a major bottleneck for serving LLMs with increasing context lengths. In response, many KV cache compression methods, such as token dropping and quantization, have been proposed. However, almost all of these methods are inherently lossy-despite minimal accuracy degradation for short outputs, their outputs increasingly diverge from full-KV-cache outputs as more tokens are decoded, which leads to catastrophic failures in code generation and tool calling. We present VeriCache, the first inference framework that ensures the same output as full-KV-cache decoding but largely preserves the high decoding throughput of a range of KV cache compression algorithms. VeriCache uses the compressed KV cache to draft tokens, then verifies them against the full KV cache. While it may seem like just speculative decoding, VeriCache requires addressing a key system challenge to work-keeping the full KV cache out of GPU memory and minimizing the overhead of swapping it in for verification. The insight is two-fold: (1) compressed-KV decoding can be parallelized with full-KV swap, because one is HBM-bandwidth-bound and the other is PCIe/network-bound, and (2) the compressed KV cache often produces output similar to the full KV cache, allowing a long drafting horizon to amortize each full-KV swap. VeriCache applies to both long-context decoding and remote prefix caching, supports a broad family of token-dropping and quantization methods through a uniform compressor interface, and composes with traditional speculative decoding. Experimental results show that VeriCache achieves up to 4X higher throughput than full-KV inference while producing identical outputs.

Linear attention has recently gained significant attention for long-context inference due to its constant decoding cost with respect to context length. However, existing serving systems typically serve linear attention by recurrently computing and updating a large linear attention state in every decoding step. Since the state is much larger than the per-token key and value, recurrent decoding incurs substantial memory access and becomes inefficient for serving linear attention. In this paper, we propose KVBuffer, an IO-aware serving mechanism for linear attention. By buffering recent keys and values, KVBuffer enables serving systems to compute linear attention outputs in more flexible and memory-efficient ways. For decoding, KVBuffer enables chunkwise computation, which reduces average memory access and decoding latency by deferring state updates and applying them in batch. For speculative decoding, KVBuffer verifies draft tokens in parallel and avoids storing temporary states. For short contexts, KVBuffer computes attention outputs directly from buffered keys and values, without creating or updating the linear attention state. We implement KVBuffer in SGLang for Qwen3-Next. Our evaluations show that KVBuffer can reduce linear attention decoding latency by up to 45.17% and increase the maximum number of serving requests by 5x for speculative decoding when verifying four draft tokens.

Large language models (LLMs) have brought autonomous agents closer to artificial general intelligence (AGI) due to their promising generalization and emergent capabilities. There is, however, a lack of studies on how LLM-based agents behave, why they could potentially fail, and how to improve them, particularly in demanding real-world planning tasks. In this paper, as an effort to fill the gap, we present our study using a realistic benchmark, TravelPlanner, where an agent must meet multiple constraints to generate accurate plans. We leverage this benchmark to address four key research questions: (1) are LLM agents robust enough to lengthy and noisy contexts when it comes to reasoning and planning? (2) can few-shot prompting adversely impact the performance of LLM agents in scenarios with long context? (3) can we rely on refinement to improve plans, and (4) can fine-tuning LLMs with both positive and negative feedback lead to further improvement? Our comprehensive experiments indicate that, firstly, LLMs often fail to attend to crucial parts of a long context, despite their ability to handle extensive reference information and few-shot examples; secondly, they still struggle with analyzing the long plans and cannot provide accurate feedback for refinement; thirdly, we propose Feedback-Aware Fine-Tuning (FAFT), which leverages both positive and negative feedback, resulting in substantial gains over Supervised Fine-Tuning (SFT). Our findings offer in-depth insights to the community on various aspects related to real-world planning applications.

Speculative decoding (SD) accelerates large language model inference by employing a faster draft model for generating multiple tokens, which are then verified in parallel by the larger target model, resulting in the text generated according to the target model distribution. However, identifying a compact draft model that is well-aligned with the target model is challenging. To tackle this issue, we propose DistillSpec that uses knowledge distillation to better align the draft model with the target model, before applying SD. DistillSpec makes two key design choices, which we demonstrate via systematic study to be crucial to improving the draft and target alignment: utilizing on-policy data generation from the draft model, and tailoring the divergence function to the task and decoding strategy. Notably, DistillSpec yields impressive 10 - 45% speedups over standard SD on a range of standard benchmarks, using both greedy and non-greedy sampling. Furthermore, we combine DistillSpec with lossy SD to achieve fine-grained control over the latency vs. task performance trade-off. Finally, in practical scenarios with models of varying sizes, first using distillation to boost the performance of the target model and then applying DistillSpec to train a well-aligned draft model can reduce decoding latency by 6-10x with minimal performance drop, compared to standard decoding without distillation.

Speculative Decoding has gained popularity as an effective technique for accelerating the auto-regressive inference process of Large Language Models (LLMs). However, Speculative Decoding entirely relies on the availability of efficient draft models, which are often lacking for many existing language models due to a stringent constraint of vocabulary incompatibility. In this work we introduce FastDraft, a novel and efficient approach for pre-training and aligning a draft model to any large language model by incorporating efficient pre-training, followed by fine-tuning over synthetic datasets generated by the target model. We demonstrate FastDraft by training two highly parameter efficient drafts for the popular Phi-3-mini and Llama-3.1-8B models. Using FastDraft, we were able to produce a draft with approximately 10 billion tokens on a single server with 8 Intel^circledR Gaudi^circledR 2 accelerators in under 24 hours. Our results show that the draft model achieves impressive results in key metrics of acceptance rate, block efficiency and up to 3x memory bound speed up when evaluated on code completion and up to 2x in summarization, text completion and instruction tasks. We validate our theoretical findings through benchmarking on the latest Intel^circledR Core^{tiny TM} Ultra, achieving a wall-clock time speedup of up to 2x, indicating a significant reduction in runtime. Due to its high quality, FastDraft unlocks large language models inference on AI-PC and other edge-devices.

Transformer-based large language models (LLMs) rely on key-value (KV) caching to avoid redundant computation during autoregressive inference. While this mechanism greatly improves efficiency, the cache size grows linearly with the input sequence length, quickly becoming a bottleneck for long-context tasks. Existing solutions mitigate this problem by evicting prompt KV that are deemed unimportant, guided by estimated importance scores. Notably, a recent line of work proposes to improve eviction quality by "glimpsing into the future", in which a draft generator produces a surrogate future response approximating the target model's true response, and this surrogate is subsequently used to estimate the importance of cached KV more accurately. However, these approaches rely on computationally expensive draft generation, which introduces substantial prefilling overhead and limits their practicality in real-world deployment. To address this challenge, we propose LookaheadKV, a lightweight eviction framework that leverages the strength of surrogate future response without requiring explicit draft generation. LookaheadKV augments transformer layers with parameter-efficient modules trained to predict true importance scores with high accuracy. Our design ensures negligible runtime overhead comparable to existing inexpensive heuristics, while achieving accuracy superior to more costly approximation methods. Extensive experiments on long-context understanding benchmarks, across a wide range of models, demonstrate that our method not only outperforms recent competitive baselines in various long-context understanding tasks, but also reduces the eviction cost by up to 14.5x, leading to significantly faster time-to-first-token. Our code is available at https://github.com/SamsungLabs/LookaheadKV.

Speculative decoding accelerates language model inference by using a fast drafter to propose candidate tokens that are then verified by a larger target model. Existing theory largely studies the stochastic, distribution-preserving setting, where the goal is to exactly sample from the target distribution. In contrast, many practical systems use greedy decoding, relaxed acceptance rules, or tree-based candidate sets, where success is governed by local ranking and threshold events rather than exact distributional equality. We develop a theory for these regimes. We identify that many common acceptance criteria have rejection regions that can be characterized as lower level sets of the target distribution. For these, we characterize the exact KL divergence required for rejection yielding exact certificates and sharp margin-based bounds for strict greedy decoding, additive and multiplicative relaxed acceptance, top-(m) relaxed criteria, and entropy-thresholded acceptance. We then extend the framework to greedy tree decoding, deriving exact and margin-only certificates for when the target greedy token remains covered by the drafter's top-(m) candidates. Finally, we evaluate the resulting certificates on Qwen3 models, showing that relaxed and tree-based criteria substantially enlarge the region of certified acceptance, especially on decoding steps with low target model distribution margin. These results complement existing distribution-preserving analyses of speculative decoding by characterizing the deterministic local acceptance events common in practical inference systems.

Large Language Models (LLMs) have shown impressive capabilities in multi-step reasoning and problem-solving.Recent works introduce multi-agent reflection frameworks where multiple LLM agents critique and refine each other's outputs using reinforcement learning (RL). However, these approaches often rely on single-shot responses and lack structural diversity in reasoning exploration. In this paper, we propose DRAFT-RL, a novel framework that integrates Chain-of-Draft (CoD) reasoning into multi-agent RL training. Instead of generating single responses, each agent produces multiple drafts per query, which are then evaluated by peer agents and a learned reward model to identify the most promising trajectory. These selected drafts are used to refine future reasoning strategies through actor-critic learning.DRAFT-RL enables explicit multi-path exploration, peer-guided reflection, and reward-aligned selection, resulting in more robust and interpretable LLM agent behavior. We evaluate our method on complex reasoning tasks including code synthesis, symbolic math, and knowledge-intensive QA,demonstrating that DRAFT-RL outperforms existing reflective and RL-based agents by significant margins in both accuracy and convergence speed

Agent-compiled knowledge bases provide persistent external knowledge for large language model (LLM) agents in open-ended, knowledge-intensive downstream tasks. Yet their quality is systematically limited by incompleteness, incorrectness, and redundancy, manifested as missing evidence or cross-document links, low-confidence or imprecise claims, and ambiguous or coreference resolution issues. Such defects compound under iterative use, degrading retrieval fidelity and downstream task performance. We present DeepRefine, a general LLM-based reasoning model for agent-compiled knowledge refinement that improves the quality of any pre-constructed knowledge bases with user queries to make it more suitable for the downstream tasks. DeepRefine performs multi-turn interactions with the knowledge base and conducts abductive diagnosis over interaction history, localizes likely defects, and executes targeted refinement actions for incremental knowledge base updates. To optimize refinement policies of DeepRefine without gold references, we introduce a Gain-Beyond-Draft (GBD) reward and train the reasoning process end-to-end via reinforcement learning. Extensive experiments demonstrate consistent downstream gains over strong baselines.

Block diffusion speculative decoding accelerates LLM inference by predicting all tokens within a block simultaneously for the target model to verify in parallel. Predicting an entire block at once requires a sufficiently capable draft model and effective utilization of the target model's internal knowledge. However, the state-of-the-art method DFlash constrains all draft layers to share a single fused representation derived from only a few target layers, limiting per-layer expressiveness and hindering further scaling of draft capacity. In this paper, we present \modelname, which flares out the narrow conditioning bottleneck of DFlash through a lightweight layer-wise fusion mechanism: each draft layer attends to its own learnable combination of a broad set of target layers at negligible overhead, simultaneously injecting richer target knowledge and providing every draft layer with a distinct input. This enhanced per-layer expressiveness enables scaling the draft model to deeper architectures with consistent gains. We further scale training data from 800K to 2.4M samples to fully exploit the enlarged capacity. On six benchmarks spanning mathematical reasoning, code generation, and conversation, \modelname attains average wall-clock speedups of 5.52x on Qwen3-4B, 5.46x on Qwen3-8B, and 3.91x on GPT-OSS-20B, improving over DFlash by roughly 11\%, 8\%, and 5\% respectively. Our code is available at https://github.com/Tencent/AngelSlim.

Retrieval-Augmented Generation (RAG) leverages retrieval tools to access external databases, thereby enhancing the generation quality of large language models (LLMs) through optimized context. However, the existing retrieval methods are constrained inherently, as they can only perform relevance matching between explicitly stated queries and well-formed knowledge, but unable to handle tasks involving ambiguous information needs or unstructured knowledge. Consequently, existing RAG systems are primarily effective for straightforward question-answering tasks. In this work, we propose MemoRAG, a novel retrieval-augmented generation paradigm empowered by long-term memory. MemoRAG adopts a dual-system architecture. On the one hand, it employs a light but long-range LLM to form the global memory of database. Once a task is presented, it generates draft answers, cluing the retrieval tools to locate useful information within the database. On the other hand, it leverages an expensive but expressive LLM, which generates the ultimate answer based on the retrieved information. Building on this general framework, we further optimize MemoRAG's performance by enhancing its cluing mechanism and memorization capacity. In our experiment, MemoRAG achieves superior performance across a variety of evaluation tasks, including both complex ones where conventional RAG fails and straightforward ones where RAG is commonly applied.

Large Language Models (LLMs) are increasingly being deployed on edge devices for long-context settings, creating a growing need for fast and efficient long-context inference. In these scenarios, the Key-Value (KV) cache is the primary bottleneck in terms of both GPU memory and latency, as the full KV cache must be loaded for each decoding step. While speculative decoding is a widely accepted technique to accelerate autoregressive decoding, existing methods often struggle to achieve significant speedups due to inefficient KV cache optimization strategies and result in low acceptance rates. To address these challenges, we propose a novel self-speculative decoding framework, QuantSpec, where the draft model shares the architecture of the target model but employs a hierarchical 4-bit quantized KV cache and 4-bit quantized weights for acceleration. QuantSpec maintains high acceptance rates (>90%) and reliably provides consistent end-to-end speedups upto sim2.5times, outperforming other self-speculative decoding methods that use sparse KV cache for long-context LLM inference. QuantSpec also reduces the memory requirements by sim 1.3times compared to these alternatives.

Deploying Large Language Models to data-scarce programming domains poses significant challenges, particularly for kernel synthesis on emerging Domain-Specific Architectures where a "Data Wall" limits available training data. While models excel on data-rich platforms like CUDA, they suffer catastrophic performance drops on data-scarce ecosystems such as NPU programming. To overcome this cold-start barrier without expensive fine-tuning, we introduce EvoKernel, a self-evolving agentic framework that automates the lifecycle of kernel synthesis from initial drafting to continual refining. EvoKernel addresses this by formulating the synthesis process as a memory-based reinforcement learning task. Through a novel value-driven retrieval mechanism, it learns stage-specific Q-values that prioritize experiences based on their contribution to the current objective, whether bootstrapping a feasible draft or iteratively refining latency. Furthermore, by enabling cross-task memory sharing, the agent generalizes insights from simple to complex operators. By building an NPU variant of KernelBench and evaluating on it, EvoKernel improves frontier models' correctness from 11.0% to 83.0% and achieves a median speedup of 3.60x over initial drafts through iterative refinement. This demonstrates that value-guided experience accumulation allows general-purpose models to master the kernel synthesis task on niche hardware ecosystems. Our official page is available at https://evokernel.zhuo.li.

Large language models (LLMs) exhibit exceptional performance across a wide range of tasks; however, their token-by-token autoregressive generation process significantly hinders inference speed. Speculative decoding presents a promising draft-then-verify framework that reduces generation latency while maintaining output distribution fidelity. Nevertheless, the draft model introduces additional computational overhead, becoming a performance bottleneck and increasing the time to first token (TTFT). Previous approaches to mitigate draft model overhead have primarily relied on heuristics and generally failed to match the quality of the draft language models. To address these challenges, we propose DuoDecoding, a novel approach that strategically deploys the draft and target models on the CPU and GPU respectively, enabling parallel decoding while preserving draft quality. Our method incorporates a hardware-aware optimal draft budget to minimize idle times and employs dynamic multi-sequence drafting to enhance draft quality. Extensive experiments across seven tasks show that DuoDecoding achieves up to 2.61x speedup in generation latency, while reducing TTFT to 83% of that in conventional speculative decoding. The Code is available at https://github.com/KaiLv69/DuoDecoding.

As LLMs are increasingly used to draft public-facing arguments, they may flatten public debate by repeatedly introducing the same polished, plausible arguments. We study argument collapse, the tendency of essays generated by different LLMs to converge to a smaller set of main arguments, sub-arguments, and paragraph-level structures. We compare 1,039 human responses from 195 New York Times (NYT) debates, 448 human responses from 61 longer-form Boston Review (BR) forums, and 23,384 LLM-generated essays. In the NYT corpus, 65.3% of human main arguments are unique within a debate, compared to 3.4% of LLM main arguments. Asking LLMs to generate diverse answers adds variation, but a typical model recovers only about half of the distinct human main arguments, with much of the added variation falling outside the observed human argument space. Collapse also appears in sub-arguments, where among essays with the same main argument, 41.0% of human sub-arguments are unique versus 9.1% from LLM responses. Qualitatively, LLMs often reuse generalized and hedged sub-arguments, while humans prefer more concrete and topic-specific ones. Structure-wise, LLM-generated essays tend to follow a more fixed arc, often opening with a direct claim and moving quickly toward proposals. The same patterns hold in longer BR essays, suggesting that argument collapse extends beyond short-form responses.

Deployment of autoregressive large language models (LLMs) is costly, and as these models increase in size, the associated costs will become even more considerable. Consequently, different methods have been proposed to accelerate the token generation process and reduce costs. Speculative decoding (SD) is among the most promising approaches to speed up the LLM decoding process by verifying multiple tokens in parallel and using an auxiliary smaller draft model to generate the possible tokens. In SD, usually, one draft model is used to serve a specific target model; however, in practice, LLMs are diverse, and we might need to deal with many target models or more than one target model simultaneously. In this scenario, it is not clear which draft model should be used for which target model, and searching among different draft models or training customized draft models can further increase deployment costs. In this paper, we first introduce a novel multi-target scenario for the deployment of draft models for faster inference. Then, we present a novel, more efficient sorted speculative decoding mechanism that outperforms regular baselines in multi-target settings. We evaluated our method on Spec-Bench in different settings, including base models such as Vicuna 7B, 13B, and LLama Chat 70B. Our results suggest that our draft models perform better than baselines for multiple target models at the same time.

Speculative decoding reduces the inference latency of a target large language model via utilizing a smaller and faster draft model. Its performance depends on a hyperparameter K -- the candidate length, i.e., the number of candidate tokens for the target model to verify in each round. However, previous methods often use simple heuristics to choose K, which may result in sub-optimal performance. We study the choice of the candidate length K and formulate it as a Markov Decision Process. We theoretically show that the optimal policy of this Markov decision process takes the form of a threshold policy, i.e., the current speculation should stop and be verified when the probability of getting a rejection exceeds a threshold value. Motivated by this theory, we propose SpecDec++, an enhanced version of speculative decoding that adaptively determines the candidate length on the fly. We augment the draft model with a trained acceptance prediction head to predict the conditional acceptance probability of the candidate tokens. SpecDec++ will stop the current speculation when the predicted probability that at least one token gets rejected exceeds a threshold. We implement SpecDec++ and apply it to the llama-2-chat 7B & 70B model pair. Our adaptive method achieves a 2.04x speedup on the Alpaca dataset (an additional 7.2% improvement over the baseline speculative decoding). On the GSM8K and HumanEval datasets, our method achieves a 2.26x speedup (9.4% improvement) and 2.23x speedup (11.1% improvement), respectively.

Speculative decoding generally dictates having a small, efficient draft model that is either pretrained or distilled offline to a particular target model series, for instance, Llama or Qwen models. However, within online deployment settings, there are two major challenges: 1) usage of a target model that is incompatible with the draft model; 2) expectation of latency improvements over usage and time. In this work, we propose OmniDraft, a unified framework that enables a single draft model to operate with any target model and adapt dynamically to user data. We introduce an online n-gram cache with hybrid distillation fine-tuning to address the cross-vocabulary mismatch across draft and target models; and further improve decoding speed by leveraging adaptive drafting techniques. OmniDraft is particularly suitable for on-device LLM applications where model cost, efficiency and user customization are the major points of contention. This further highlights the need to tackle the above challenges and motivates the ``one drafter for all'' paradigm. We showcase the proficiency of the OmniDraft framework by performing online learning on math reasoning, coding and text generation tasks. Notably, OmniDraft enables a single Llama-68M model to pair with various target models including Vicuna-7B, Qwen2-7B and Llama3-8B models for speculative decoding; and additionally provides up to 1.5-2x speedup.

Speculative decoding is a widely adopted technique for accelerating inference in large language models (LLMs), yet its application to vision-language models (VLMs) remains underexplored, with existing methods achieving only modest speedups (<1.5x). This gap is increasingly significant as multimodal capabilities become central to large-scale models. We hypothesize that large VLMs can effectively filter redundant image information layer by layer without compromising textual comprehension, whereas smaller draft models struggle to do so. To address this, we introduce Vision-Aware Speculative Decoding (ViSpec), a novel framework tailored for VLMs. ViSpec employs a lightweight vision adaptor module to compress image tokens into a compact representation, which is seamlessly integrated into the draft model's attention mechanism while preserving original image positional information. Additionally, we extract a global feature vector for each input image and augment all subsequent text tokens with this feature to enhance multimodal coherence. To overcome the scarcity of multimodal datasets with long assistant responses, we curate a specialized training dataset by repurposing existing datasets and generating extended outputs using the target VLM with modified prompts. Our training strategy mitigates the risk of the draft model exploiting direct access to the target model's hidden states, which could otherwise lead to shortcut learning when training solely on target model outputs. Extensive experiments validate ViSpec, achieving, to our knowledge, the first substantial speedup in VLM speculative decoding. Code is available at https://github.com/KangJialiang/ViSpec.

Large language models are limited in deployment by GPU memory and inference latency. We present Minima, a production compression pipeline that learns where and how to structurally compress a Transformer and turns that compression into real serving gains. Minima trains a lightweight convolutional predictor to estimate layer- and patch-level sensitivity, applies a mixture of Tucker, tensor-train, and tensor-ring decompositions to low-sensitivity regions, performs a short healing fine-tune, and executes the resulting operators with custom Triton and CUDA kernels. The reduced memory footprint enables speculative decoding with a small draft model and a larger verifier. On Qwen3-32B at an 8k-token context window, Minima reduces peak VRAM from 64 GiB to 40 GiB. For a single active request, throughput increases from 40 tokens per second (baseline) to 50 tokens per second (Minima) and 75 tokens per second (Minima with speculative decoding). Under 50 parallel requests, throughput is 34, 44, and 53 tokens per second respectively, showing that Minima remains effective under high concurrency even when speculative decoding gains compress. We position Minima relative to recent tensor-network, low-rank plus quantization, and cross-layer sharing methods, and argue that it is a practical step toward more aggressive structural compression via shared tensor backbones with tiny per-layer adapters.

Reinforcement learning (RL) has become a representative post-training paradigm for LLMs, enabling strong reasoning and agentic capabilities. However, rollout generation remains a dominant latency bottleneck because autoregressive sampling decodes responses sequentially and a small number of long-tailed generations often determine completion time. Speculative decoding (SD) offers a natural way to address this bottleneck, as it is a well-established technique for serving fixed LLMs that reduces latency by rapidly drafting tokens and accepting them through parallel verification while preserving the target-model distribution. However, its practical speedups do not directly carry over to RL rollouts: (i) the evolving target policy makes any fixed drafter increasingly mismatched with the policy's output distribution; and (ii) active batch sizes shrink throughout rollout decoding, shifting decoding from compute-bound to memory-bound regimes where parallel verification can exploit underutilized compute. Therefore, accelerating RL rollouts requires both a drafter that remains effective under long, high-temperature generations from an evolving policy and system-aware use of SD that avoids compute-bound regimes. We present EfficientRollout, a system-aware self-SD framework designed to address this gap for RL rollouts. EfficientRollout induces a quantized drafter from the target model (i.e. self-speculative decoding), keeping it coupled to the evolving policy without separate drafter pretraining or online adaptation. It further coordinates a system-aware SD toggle policy with acceptance-aware draft-length adaptation, enabling speculation only in beneficial regimes while matching the drafting budget to evolving drafter quality. EfficientRollout reduces rollout and end-to-end latency by up to 19.6% and 12.7%, respectively, over an accelerated AR rollout baseline, while preserving final model quality.

Autoregressive decoding is bottlenecked by its sequential nature. Speculative decoding has become a standard way to accelerate inference by using a fast draft model to predict upcoming tokens from a slower target model, and then verifying them in parallel with a single target model forward pass. However, speculative decoding itself relies on a sequential dependence between speculation and verification. We introduce speculative speculative decoding (SSD) to parallelize these operations. While a verification is ongoing, the draft model predicts likely verification outcomes and prepares speculations pre-emptively for them. If the actual verification outcome is then in the predicted set, a speculation can be returned immediately, eliminating drafting overhead entirely. We identify three key challenges presented by speculative speculative decoding, and suggest principled methods to solve each. The result is Saguaro, an optimized SSD algorithm. Our implementation is up to 2x faster than optimized speculative decoding baselines and up to 5x faster than autoregressive decoding with open source inference engines.

Speculative decoding accelerates LLM inference by using a draft model to look ahead, but gains are capped by the cost of autoregressive draft generation: increasing draft size elevates acceptance rates but introduces additional latency overhead exacerbating the speed-accuracy tradeoff. Prior methods (Medusa, Hydra, EAGLE) partially reduce draft cost but either degrade acceptance or introduce overheads that limit scaling. We present Mirror Speculative Decoding (Mirror-SD), an inference algorithm that breaks the latency-acceptance tradeoff. Mirror-SD launches branch-complete rollouts from early-exit signals in parallel with the target model's suffix and explicitly maps computation across heterogeneous accelerators (GPU and NPU) to exploit cross-device parallelism. The draft speculates forward continuations for the target to verify, while the target simultaneously speculates correction paths for the draft, converting speculation into two complementary execution pipelines. To further cut draft latency without weakening acceptance semantics, we add speculative streaming so the draft emits multiple tokens per step. This dual strategy of parallel heterogeneous execution plus multi-token speculative streaming pushes speculative decoding toward its ideal regime of high acceptance with low overhead. On SpecBench with server-scale models from 14B to 66B parameters, Mirror-SD delivers consistent end-to-end gains, achieving 2.8x-5.8x wall-time speedups across diverse tasks and a 30% average relative improvement over the strongest baseline, EAGLE3.

Machine learning (ML) has become crucial in modern life, with growing interest from researchers and the public. Despite its potential, a significant entry barrier prevents widespread adoption, making it challenging for non-experts to understand and implement ML techniques. The increasing desire to leverage ML is counterbalanced by its technical complexity, creating a gap between potential and practical application. This work introduces asanAI, an offline-first, open-source, no-code machine learning toolkit designed for users of all skill levels. It allows individuals to design, debug, train, and test ML models directly in a web browser, eliminating the need for software installations and coding. The toolkit runs on any device with a modern web browser, including smartphones, and ensures user privacy through local computations while utilizing WebGL for enhanced GPU performance. Users can quickly experiment with neural networks and train custom models using various data sources, supported by intuitive visualizations of network structures and data flows. asanAI simplifies the teaching of ML concepts in educational settings and is released under an open-source MIT license, encouraging modifications. It also supports exporting models in industry-ready formats, empowering a diverse range of users to effectively learn and apply machine learning in their projects. The proposed toolkit is successfully utilized by researchers of ScaDS.AI to swiftly draft and test machine learning ideas, by trainers to effectively educate enthusiasts, and by teachers to introduce contemporary ML topics in classrooms with minimal effort and high clarity.

Creating visual layouts is a critical step in graphic design. Automatic generation of such layouts is essential for scalable and diverse visual designs. To advance conditional layout generation, we introduce BLT, a bidirectional layout transformer. BLT differs from previous work on transformers in adopting non-autoregressive transformers. In training, BLT learns to predict the masked attributes by attending to surrounding attributes in two directions. During inference, BLT first generates a draft layout from the input and then iteratively refines it into a high-quality layout by masking out low-confident attributes. The masks generated in both training and inference are controlled by a new hierarchical sampling policy. We verify the proposed model on six benchmarks of diverse design tasks. Experimental results demonstrate two benefits compared to the state-of-the-art layout transformer models. First, our model empowers layout transformers to fulfill controllable layout generation. Second, it achieves up to 10x speedup in generating a layout at inference time than the layout transformer baseline. Code is released at https://shawnkx.github.io/blt.

Following each patient visit, physicians draft long semi-structured clinical summaries called SOAP notes. While invaluable to clinicians and researchers, creating digital SOAP notes is burdensome, contributing to physician burnout. In this paper, we introduce the first complete pipelines to leverage deep summarization models to generate these notes based on transcripts of conversations between physicians and patients. After exploring a spectrum of methods across the extractive-abstractive spectrum, we propose Cluster2Sent, an algorithm that (i) extracts important utterances relevant to each summary section; (ii) clusters together related utterances; and then (iii) generates one summary sentence per cluster. Cluster2Sent outperforms its purely abstractive counterpart by 8 ROUGE-1 points, and produces significantly more factual and coherent sentences as assessed by expert human evaluators. For reproducibility, we demonstrate similar benefits on the publicly available AMI dataset. Our results speak to the benefits of structuring summaries into sections and annotating supporting evidence when constructing summarization corpora.

Speculative decoding has become the standard approach for accelerating Large Language Model (LLM) inference. It exploits a lossless draft-then-verify procedure to circumvent the latency of autoregressive decoding, achieving impressive speed-ups. Yet, current speculative decoding approaches remain limited by two fundamental bottlenecks: (1) the autoregressive dependency during drafting which limits parallelism, and (2) frequent rejections of draft tokens caused by misalignment between the draft and verify models. This paper proposes SpecDiff-2, a novel framework to jointly address these two bottlenecks. It leverages discrete diffusion as a non-autoregressive drafter to address bottleneck (1) and develops novel techniques to calibrate discrete diffusion drafters with autoregressive verifiers, addressing bottleneck (2). Experimental results across a comprehensive benchmark suite show that SpecDiff-2 achieves a new state-of-the-art across reasoning, coding, and mathematical benchmarks, improving tokens-per-second by up to an average of +55% over previous baselines and obtaining up to 5.5x average speed-up over standard decoding, without any loss of accuracy.

Speculative decoding accelerates autoregressive large language model (LLM) inference by using a lightweight draft model to propose candidate tokens that are then verified in parallel by the target model. The speedup is significantly determined by the acceptance rate, yet standard training minimizes Kullback-Leibler (KL) divergence as a proxy objective. While KL divergence and acceptance rate share the same global optimum, small draft models, having limited capacity, typically converge to suboptimal solutions where minimizing KL does not guarantee maximizing acceptance rate. To address this issue, we propose LK losses, special training objectives that directly target acceptance rate. Comprehensive experiments across four draft architectures and six target models, ranging from 8B to 685B parameters, demonstrate consistent improvements in acceptance metrics across all configurations compared to the standard KL-based training. We evaluate our approach on general, coding and math domains and report gains of up to 8-10% in average acceptance length. LK losses are easy to implement, introduce no computational overhead and can be directly integrated into any existing speculator training framework, making them a compelling alternative to the existing draft training objectives.

Speculative decoding is a widely adopted technique for accelerating inference in large language models (LLMs), but its performance degrades on long inputs due to increased attention cost and reduced draft accuracy. We introduce SpecExtend, a drop-in enhancement that improves the performance of speculative decoding on long sequences without any additional training. SpecExtend integrates efficient attention mechanisms such as FlashAttention and Hybrid Tree Attention into both the draft and target models, reducing latency across all stages. To improve draft accuracy and speed, we propose Cross-model Retrieval, a novel KV cache update strategy that uses the target model's attention scores to dynamically select relevant context for the draft model. Extensive evaluations on three long-context understanding datasets show that SpecExtend accelerates standard tree-based speculative decoding by up to 2.22x for inputs up to 16K tokens, providing an effective solution for speculative decoding of long sequences. The code is available at https://github.com/jycha98/SpecExtend .

Large Language Models (LLMs) have achieved remarkable success in natural language understanding and generation, but their deployment is constrained by high computational demands. Deploying smaller LLMs directly on the edge can circumvent this, but with degraded accuracy. Deploying smaller cloud-based big LLMs preserves performance, but at the cost of expensive per-token computation. We present a distributed inference framework, , that integrates speculative decoding (SD) across edge and cloud. A compact draft model deployed on the edge generates candidate tokens rapidly, and a large verifier model on the cloud validates these tokens in parallel. Accepted tokens are retained, while only rejections trigger verifier correction, substantially reducing the number of cloud queries. Our plug-and-play design shifts the bulk of computation to the edge, significantly lowers inference time and cloud cost, and preserves the accuracy of the big model without any retraining requirement. Our approach demonstrates a practical path toward scalable, cost-efficient, and accurate deployment of LLMs in real-world environments. Experimental results across multiple Natural Language Processing tasks using SpecBench and CNN/Dailymail datasets demonstrate that reduces the cloud model calls by 76% with zero loss in accuracy as compared to the full model.

SmartAvatar is a vision-language-agent-driven framework for generating fully rigged, animation-ready 3D human avatars from a single photo or textual prompt. While diffusion-based methods have made progress in general 3D object generation, they continue to struggle with precise control over human identity, body shape, and animation readiness. In contrast, SmartAvatar leverages the commonsense reasoning capabilities of large vision-language models (VLMs) in combination with off-the-shelf parametric human generators to deliver high-quality, customizable avatars. A key innovation is an autonomous verification loop, where the agent renders draft avatars, evaluates facial similarity, anatomical plausibility, and prompt alignment, and iteratively adjusts generation parameters for convergence. This interactive, AI-guided refinement process promotes fine-grained control over both facial and body features, enabling users to iteratively refine their avatars via natural-language conversations. Unlike diffusion models that rely on static pre-trained datasets and offer limited flexibility, SmartAvatar brings users into the modeling loop and ensures continuous improvement through an LLM-driven procedural generation and verification system. The generated avatars are fully rigged and support pose manipulation with consistent identity and appearance, making them suitable for downstream animation and interactive applications. Quantitative benchmarks and user studies demonstrate that SmartAvatar outperforms recent text- and image-driven avatar generation systems in terms of reconstructed mesh quality, identity fidelity, attribute accuracy, and animation readiness, making it a versatile tool for realistic, customizable avatar creation on consumer-grade hardware.

Speculative decoding has emerged as a widely adopted method to accelerate large language model inference without sacrificing the quality of the model outputs. While this technique has facilitated notable speed improvements by enabling parallel sequence verification, its efficiency remains inherently limited by the reliance on incremental token generation in existing draft models. To overcome this limitation, this paper proposes an adaptation of speculative decoding which uses discrete diffusion models to generate draft sequences. This allows parallelization of both the drafting and verification steps, providing significant speed-ups to the inference process. Our proposed approach, Speculative Diffusion Decoding (SpecDiff), is validated on standard language generation benchmarks and empirically demonstrated to provide a up to 8.7x speed-up over standard generation processes and up to 2.5x speed-up over existing speculative decoding approaches.

AI-assisted research is crossing a threshold: fully automated systems can now generate research papers for as little as $15, while long-horizon agents can execute experiments, draft manuscripts, and simulate critique with minimal human input. Yet this productivity frontier exposes a deeper integrity problem: under scientific pressure, even frontier LLMs still fabricate results, miss hidden errors, and fail to judge novelty reliably. Studying developments through April 2026, we present an end-to-end analysis of AI across the complete research lifecycle, organized into four epistemological phases: Creation (idea generation, literature review, coding & experiments, tables & figures), Writing (paper writing), Validation (peer review, rebuttal & revision), and Dissemination (posters, slides, videos, social media, project pages, and interactive agents). We identify a sharp, stage-dependent boundary between reliable assistance and unreliable autonomy: AI excels at structured, retrieval-grounded, and tool-mediated tasks, but remains fragile for genuinely novel ideas, research-level experiments, and scientific judgment. Generated ideas often degrade after implementation, research code lags far behind pattern-matching benchmarks, and end-to-end autonomous systems have not yet consistently reached major-venue acceptance standards. We further show that greater automation can obscure rather than eliminate failure modes, making human-governed collaboration the most credible deployment paradigm. Finally, we provide a structured taxonomy, benchmark suite, and tool inventory, cross-stage design principles, and a practitioner-oriented playbook, with resources maintained at our project page.

Large language models (LLMs) have revolutionized natural language processing and broadened their applicability across diverse commercial applications. However, the deployment of these models is constrained by high inference time in multilingual settings. To mitigate this challenge, this paper explores a training recipe of an assistant model in speculative decoding, which are leveraged to draft and-then its future tokens are verified by the target LLM. We show that language-specific draft models, optimized through a targeted pretrain-and-finetune strategy, substantially brings a speedup of inference time compared to the previous methods. We validate these models across various languages in inference time, out-of-domain speedup, and GPT-4o evaluation.

Speculative decoding greatly increases the interactivity of autoregressive language models by trading off computation for extra tokens generated in a single forward pass. Factorized draft models are especially efficient because they predict future-token marginals in parallel, but their independence assumption causes acceptance rates to degrade sharply as the speculative budget grows. We analyze this limitation and introduce Weaver, a lightweight autoregressive adapter that constructs proposal trees from the top-K marginals of a factorized drafter. Weaver restores conditional dependencies between proposed tokens while avoiding a full-vocabulary projection. To support fast verification for models with Gated Delta Net layers, we derive a rollback-free tree-verification algorithm and implement optimized CUDA kernels in SGLang. By combining these model and systems contributions we achieve a 4.37-fold speedup over autoregressive decoding, and outperform a highly optimized DFlash baseline by 24.7%.

Recent byte-level language models (LMs) match the performance of token-level models without relying on subword vocabularies, yet their utility is limited by slow, byte-by-byte autoregressive generation. We address this bottleneck in the Byte Latent Transformer (BLT) through new training and generation techniques. First, we introduce BLT Diffusion (BLT-D), a new model and our fastest BLT variant, trained with an auxiliary block-wise diffusion objective alongside the standard next-byte prediction loss. This enables an inference procedure that generates multiple bytes in parallel per decoding step, substantially reducing the number of forward passes required to generate a sequence. Second, we propose two extensions inspired by speculative decoding that trade some of this speed for higher generation quality: BLT Self-speculation (BLT-S), in which BLT's local decoder continues generating past its normal patch boundaries to draft bytes, which are then verified with a single full-model forward pass; and BLT Diffusion+Verification (BLT-DV), which augments BLT-D with an autoregressive verification step after diffusion-based generation. All methods may achieve an estimated memory-bandwidth cost over 50% lower than BLT on generation tasks. Each approach offers its own unique advantages, together removing key barriers to the practical use of byte-level LMs.

One of the most compelling features of global discrete diffusion language models is their global bidirectional contextual capability. However, existing block-based diffusion studies tend to introduce autoregressive priors, which, while offering benefits, can cause models to lose this global coherence at the macro level. To regain global contextual understanding while preserving the advantages of the semi-autoregressive paradigm, we propose Diffusion in Diffusion, a 'draft-then-refine' framework designed to overcome the irreversibility and myopia problems inherent in block diffusion models. Our approach first employs block diffusion to generate rapid drafts using small blocks, then refines these drafts through global bidirectional diffusion with a larger bidirectional receptive field. We utilize snapshot confidence remasking to identify the most critical tokens that require modification, and apply mix-scale training to expand the block diffusion model's global capabilities. Empirical results demonstrate that our approach sets a new benchmark for discrete diffusion models on the OpenWebText dataset. Using only 26% of the fine-tuning budget of baseline models, we reduce generative perplexity from 25.7 to 21.9, significantly narrowing the performance gap with autoregressive models.

Large language models (LLMs) have the potential to transform medicine, but real-world clinical scenarios contain extraneous information that can hinder performance. The rise of assistive technologies like ambient dictation, which automatically generates draft notes from live patient encounters, has the potential to introduce additional noise making it crucial to assess the ability of LLM's to filter relevant data. To investigate this, we developed MedDistractQA, a benchmark using USMLE-style questions embedded with simulated real-world distractions. Our findings show that distracting statements (polysemous words with clinical meanings used in a non-clinical context or references to unrelated health conditions) can reduce LLM accuracy by up to 17.9%. Commonly proposed solutions to improve model performance such as retrieval-augmented generation (RAG) and medical fine-tuning did not change this effect and in some cases introduced their own confounders and further degraded performance. Our findings suggest that LLMs natively lack the logical mechanisms necessary to distinguish relevant from irrelevant clinical information, posing challenges for real-world applications. MedDistractQA and our results highlights the need for robust mitigation strategies to enhance LLM resilience to extraneous information.

This report describes our approach to the Efficient Qwen Competition, where the goal is to enable low-latency serving of Qwen3.5-4B on a resource-constrained NVIDIA A10G GPU. Our system combines a quantized target model with speculative decoding. To recover accuracy, we apply quantization-aware distillation to the target model while retaining the original quantization grid. To speed up decoding, a block-diffusion drafter specialized for the quantized target model is trained using a two-stage procedure: first learning from the high-precision target and then adapting to the low-precision target. Because the drafter is invoked at every speculative decoding step, we further reduce its overhead with quantization and sliding-window attention, preserving draft-token acceptance while improving long-context decoding latency. As a result, our submission achieves a 6.978times average speedup over the baseline while satisfying the required quality thresholds, ranking 3rd overall. We hope these results provide useful insights for practical LLM inference. The code and resources are available at https://github.com/nota-github/adaptfm-quant-dflash

Multi-path speculative decoding accelerates lossless sampling from a target model by using a cheaper draft model to generate a draft tree of tokens, and then applies a verification algorithm that accepts a subset of these. While prior work has proposed various verification algorithms for i.i.d rollouts, their relative performance under matched settings remains unclear. In this work, we firstly present a systematic evaluation of verification strategies across model families, tasks, and sampling regimes, and find that Traversal Verification dominates consistently, with OT-based methods lagging far behind. Our analysis uncovers that this occurs because OT-based methods achieve high multi-token acceptance near the root of the draft tree, while multi-token gains are most impactful deeper in the draft tree, where draft and target distributions diverge. Based on this insight, we propose delayed tree expansion, which drafts a partial single path, delaying the i.i.d. branching point. We show that delayed tree expansion preserves the target distribution and improves on root-node i.i.d rollouts. Further, we develop a dynamic neural selector that estimates the expected block efficiency of optimal-transport-based verification methods from draft and target features, enabling context-dependent expansion decisions. Our neural selector allows OT-based methods like SpecInfer to outperform Traversal Verification for the first time, achieving 5% higher average throughput across a wide range of models, datasets, and sampling settings.

We present Mellum 2, an open-weight 12B-parameter Mixture-of-Experts (MoE) language model with 2.5B active parameters per token. Mellum 2 is a general-purpose language model specialized in software engineering, spanning code generation and editing, debugging, multi-step reasoning, tool use and function calling, agentic coding, and conversational programming assistance, and it is the successor to the completion-focused 4B dense Mellum model. The architecture builds on the Mixture-of-Experts (64 experts, 8 active) and combines Grouped-Query Attention with 4 KV heads, Sliding Window Attention on three of every four layers, and a single Multi-Token Prediction head that doubles as both an auxiliary pre-training objective and a built-in draft model for speculative decoding; each choice was validated by ablation with inference efficiency on commodity GPUs as a design constraint. Pre-training spans approximately 10.6 trillion tokens through a three-phase curriculum that progressively shifts the mixture from diverse web data toward curated code and mathematical content, optimized with Muon under FP8 hybrid precision and a Warmup-Hold-Decay schedule with linear decay to zero. The pre-trained base is extended to a 128K context window via a layer-selective YaRN and then post-trained in two stages (supervised fine-tuning followed by RLVR), yielding two released variants: an Instruct model that answers directly and a Thinking model that emits an explicit reasoning trace before its final answer. Across code generation, math and reasoning, tool use, knowledge, and safety benchmarks, Mellum 2 is competitive with open-weight baselines in the 4B-14B range while running at the per-token compute of a 2.5B dense model. We release the base, instruct, and thinking checkpoints, together with this report on the architecture decisions, data pipeline, and training recipe behind them, under the Apache 2.0 license.

Speculative decoding is an effective and lossless approach for accelerating LLM inference. However, existing widely adopted model-based draft designs, such as EAGLE3, improve accuracy at the cost of multi-step autoregressive inference, resulting in high drafting latency and ultimately rendering the drafting stage itself a performance bottleneck. Inspired by diffusion-based large language models (dLLMs), we propose DART, which leverages parallel generation to reduce drafting latency. DART predicts logits for multiple future masked positions in parallel within a single forward pass based on hidden states of the target model, thereby eliminating autoregressive rollouts in the draft model while preserving a lightweight design. Based on these parallel logit predictions, we further introduce an efficient tree pruning algorithm that constructs high-quality draft token trees with N-gram-enforced semantic continuity. DART substantially reduces draft-stage overhead while preserving high draft accuracy, leading to significantly improved end-to-end decoding speed. Experimental results demonstrate that DART achieves a 2.03x--3.44x wall-clock time speedup across multiple datasets, surpassing EAGLE3 by 30% on average and offering a practical speculative decoding framework. Code is released at https://github.com/fvliang/DART.

Speculative decoding is an effective technique for accelerating large language model inference by drafting multiple tokens in parallel. In practice, its speedup is often bottlenecked by a rigid verification step that strictly enforces the accepted token distribution to exactly match the target model. This constraint leads to the rejection of many plausible tokens, lowering the acceptance rate and limiting overall time speedup. To overcome this limitation, we propose Dynamic Verification Relaxed Speculative Decoding (DIVERSED), a relaxed verification framework that improves time efficiency while preserving generation quality. DIVERSED learns an ensemble-based verifier that blends the draft and target model distributions with a task-dependent and context-dependent weight. We provide theoretical justification for our approach and demonstrate empirically that DIVERSED achieves substantially higher inference efficiency compared to standard speculative decoding methods. Code is available at: https://github.com/comeusr/diversed.

Algorithmic fairness has conventionally adopted the mathematically convenient perspective of racial color-blindness (i.e., difference unaware treatment). However, we contend that in a range of important settings, group difference awareness matters. For example, differentiating between groups may be necessary in legal contexts (e.g., the U.S. compulsory draft applies to men but not women) and harm assessments (e.g., referring to girls as ``terrorists'' may be less harmful than referring to Muslim people as such). Thus, in contrast to most fairness work, we study fairness through the perspective of treating people differently -- when it is contextually appropriate to. We first introduce an important distinction between descriptive (fact-based), normative (value-based), and correlation (association-based) benchmarks. This distinction is significant because each category requires separate interpretation and mitigation tailored to its specific characteristics. Then, we present a benchmark suite composed of eight different scenarios for a total of 16k questions that enables us to assess difference awareness. Finally, we show results across ten models that demonstrate difference awareness is a distinct dimension to fairness where existing bias mitigation strategies may backfire.

Block-diffusion drafters like dFlash generate an entire block of draft tokens in a single forward pass, drastically reducing the overhead of multiple-token drafting in speculative decoding. The crucial final step of the single-pass discrete denoising process involves using the logit distribution at each position to sample conditionally independent tokens. The resulting draft is thus a set of per-position marginals, rather than a joint distribution: no draft token is guaranteed to depend on its predecessors. Such independently sampled marginals tend to produce sequences with tokens that are individually likely, but jointly improbable under the target model's distribution, which verifies each token conditionally. This can cause early rejection and limits acceptance length. To address this, we propose xPress as a means to restore the missing causality in diffusion drafters. xPress is a lightweight causal refiner that reconciles the whole diffusion block at once through parallel refinement, restoring and propagating causal dependencies across the draft without a token-by-token loop. On Qwen3-8B, across seven math, code, and chat benchmarks, xPress raises acceptance length by about 30% on average (up to +56%) and its end-to-end decoding throughput by about 1.3 on average (up to 1.7) compared to the original dFlash diffusion drafter.

Generation of plausible yet incorrect factual information, termed hallucination, is an unsolved issue in large language models. We study the ability of language models to deliberate on the responses they give in order to correct their mistakes. We develop the Chain-of-Verification (CoVe) method whereby the model first (i) drafts an initial response; then (ii) plans verification questions to fact-check its draft; (iii) answers those questions independently so the answers are not biased by other responses; and (iv) generates its final verified response. In experiments, we show CoVe decreases hallucinations across a variety of tasks, from list-based questions from Wikidata, closed book MultiSpanQA and longform text generation.

Speculative sampling has emerged as an important technique for accelerating the auto-regressive generation process of large language models (LLMs) by utilizing a draft-then-verify mechanism to produce multiple tokens per forward pass. While state-of-the-art speculative sampling methods use only a single layer and a language modeling (LM) head as the draft model to achieve impressive layer compression, their efficiency gains are substantially reduced for large-vocabulary LLMs, such as Llama-3-8B with a vocabulary of 128k tokens. To address this, we present FR-Spec, a frequency-ranked speculative sampling framework that optimizes draft candidate selection through vocabulary space compression. By constraining the draft search to a frequency-prioritized token subset, our method reduces LM Head computation overhead by 75% while ensuring the equivalence of the final output distribution. Experiments across multiple datasets demonstrate an average of 1.12times speedup over the state-of-the-art speculative sampling method EAGLE-2.

Chain-of-thought (CoT) is a standard approach for eliciting reasoning capabilities from large language models (LLMs). However, the common CoT paradigm treats thinking as a prerequisite for answering, which can delay access to plausible answers and incur unnecessary token costs even when the model is able to identify an answer before extended thinking, a behavior known as performative reasoning. In this paper, we introduce CopT, a reformulated reasoning pipeline that reverses the usual order of thinking and answering. Instead of thinking before answering, CopT first elicits a draft answer and then invokes subsequent on-policy thinking conditioned on its own draft answer for reflection and correction. To assess whether the draft answer should be trusted, CopT recasts continuous embeddings as inference-time contrastive verifiers. Specifically, it contrasts the model's support for the same generated tokens under discrete-token inputs and continuous-embedding inputs, yielding a sequence-level reverse KL estimator for answer reliability. Our analysis shows that under certain assumptions, the expected estimate equals the mutual information between the unresolved latent state and the emitted answer token, explaining why it captures answer-relevant uncertainty rather than arbitrary uncertainty in the latent state. When the answer is deemed insufficiently reliable, CopT performs further on-policy thinking, where a second KL estimator dynamically controls draft-answer visibility, preserving useful partial information while reducing the risk of being misled by unreliable content. Across mathematics, coding, and agentic reasoning tasks, CopT improves peak accuracy by up to 23% and reduces token usage by up to 57% at comparable or higher accuracy, without any additional training. The code is available at https://github.com/sdc17/CopT.

Speculative decoding (SD) is a powerful technique for accelerating the inference process of large language models (LLMs) without sacrificing accuracy. Typically, SD employs a small draft model to generate a fixed number of draft tokens, which are then verified in parallel by the target model. However, our experiments reveal that the optimal draft length varies significantly across different decoding steps. This variation suggests that using a fixed draft length limits the potential for further improvements in decoding speed. To address this challenge, we propose Pacer, a novel approach that dynamically controls draft length using a lightweight, trainable pre-verification layer. This layer pre-verifies draft tokens blockwise before they are sent to the target model, allowing the draft model to stop token generation if the blockwise pre-verification fails. We implement Pacer on multiple SD model pairs and evaluate its performance across various benchmarks. Our results demonstrate that Pacer achieves up to 2.66x Speedup over autoregressive decoding and consistently outperforms standard speculative decoding. Furthermore, when integrated with Ouroboros, Pacer attains up to 3.09x Speedup.

Bug reports serve as task specifications for repository-level automated program repair (APR) agents, but they often describe only the observed failure and omit repair-relevant information such as the failure-inducing behavior, behavioral requirement, and implementation scope. As a result, a repair agent may inspect irrelevant code, infer an incorrect requirement, or generate a patch that addresses the reported symptom without restoring the intended repository behavior. We present TrajSpec, a trajectory-guided approach for repository-supported bug report specification refinement. Given an original report and a pre-fix repository, TrajSpec runs a trajectory-collection agent and uses the resulting unverified trajectory as a source of trajectory-derived specification evidence. It organizes this evidence into a three-level representation consisting of a high-level interpretation of the issue, diagnostic findings supporting that interpretation, and concrete repository observations. TrajSpec then generates a draft refined report and applies repository-based review to remove unsupported claims, revise uncertain claims, and add repository-supported details. We evaluate TrajSpec on all 300 SWE-Bench Lite instances using Mini-SWE-Agent V2. TrajSpec's refined reports improve Pass@1 from 41.00% to 59.67% with GPT-5-mini and from 54.67% to 64.33% with MiniMax M2.5. On a stratified sample of 100 instances, TrajSpec's refined reports also improve Pass@1 from 41.00% to 71.00% with Agentless and from 47.00% to 72.00% with AutoCodeRover. Ablation results show that removing repository-based review or the hierarchical evidence representation reduces Pass@1 from 59.67% to 48.00% and 47.67%, respectively. Overall, TrajSpec provides actionable repository-supported context that consistently improves repair performance.

Speculative decoding accelerates memory-bound LLM inference without quality degradation by using a fast drafter to propose multiple candidate tokens and the target model to verify them in parallel. However, conventional sequential speculative decoding suffers from mutual waiting between drafting and verification, and repeated exchange of intermediate states further increases memory access overhead. Parallel speculative decoding addresses this limitation by performing drafting and verification within a single target forward pass, allowing future drafts to be prepared while current candidates are being verified. Although effective at small batch sizes, existing parallel speculative decoding methods either require costly continual pretraining with quality degradation or suffer from low acceptance rates. More importantly, this paradigm inherently suffers from uncertainty in both the bonus token and the accepted length, leading to draft verification mismatch and causing throughput gains to collapse at large batch sizes. To address these limitations, we introduce FlexDraft, a lossless speculative decoding framework that flexibly adapts to varying batch sizes through three key designs. (1) Attention Tuning enables block diffusion drafting by tuning only the attention projectors of the final few layers on mask tokens, while keeping the autoregressive path frozen to preserve the target distribution and produce high quality drafts with minimal trainable parameters. (2) Bonus-guided Calibration uses a lightweight MLP conditioned on the resolved bonus token to calibrate draft logits, mitigating draft verification mismatch caused by bonus token uncertainty. (3) Flex Decoding dynamically switches between parallel draft and verify at small batch sizes and sequential draft then verify at large batch sizes, and adjusts verification length based on draft confidence to eliminate redundant computation.

Growing demands from tasks like code generation, deep reasoning, and long-document understanding have made long-context generation a crucial capability for large language models (LLMs). Speculative decoding is one of the most direct and effective approaches for accelerating generation. It follows a draft-verify paradigm, where a lightweight draft model proposes several candidate tokens and the target model verifies them. However, we find that as the context length grows, verification becomes the dominant bottleneck. To further accelerate speculative decoding in long-context generation, we introduce SpecPV, a self-speculative decoding approach that performs fast verification using partial key-value states (KV) and periodically applies full verification to eliminate accumulated errors. We validate SpecPV across multiple long-context benchmarks and models, including LLaMA-3.1-8B-Instruct and Qwen3-series. Experimental results show that SpecPV achieves up to 6x decoding speedup over standard autoregressive decoding with minor degradation.

Speculative decoding accelerates inference in large language models (LLMs) by generating multiple draft tokens simultaneously. However, existing methods often struggle with token misalignment between the training and decoding phases, limiting their performance. To address this, we propose GRIFFIN, a novel framework that incorporates a token-alignable training strategy and a token-alignable draft model to mitigate misalignment. The training strategy employs a loss masking mechanism to exclude highly misaligned tokens during training, preventing them from negatively impacting the draft model's optimization. The token-alignable draft model introduces input tokens to correct inconsistencies in generated features. Experiments on LLaMA-series and Vicuna models demonstrate that GRIFFIN achieves an average acceptance length improvement of over 7\% and a speedup ratio exceeding 8%, outperforming current SoTAs as shown in Fig. 1 (a) and (b).

Efficient inference in large language models (LLMs) has become a critical focus as their scale and complexity grow. Traditional autoregressive decoding, while effective, suffers from computational inefficiencies due to its sequential token generation process. Speculative decoding addresses this bottleneck by introducing a two-stage framework: drafting and verification. A smaller, efficient model generates a preliminary draft, which is then refined by a larger, more sophisticated model. This paper provides a comprehensive survey of speculative decoding methods, categorizing them into draft-centric and model-centric approaches. We discuss key ideas associated with each method, highlighting their potential for scaling LLM inference. This survey aims to guide future research in optimizing speculative decoding and its integration into real-world LLM applications.

Speculative decoding accelerates autoregressive generation by letting a lightweight draft model propose future tokens that a larger target model then verifies in parallel. In practice, however, draft models are usually trained on broad generic corpora, which leaves it unclear how much speculative decoding quality depends on the draft training distribution. We study this question with lightweight HASS and EAGLE-2 drafters trained on MathInstruct, ShareGPT, and mixed-data variants, evaluated on MT-Bench, GSM8K, MATH-500, and SVAMP. Measured by acceptance length, task-specific training yields clear specialization: MathInstruct-trained drafts are strongest on reasoning benchmarks, while ShareGPT-trained drafts are strongest on MT-Bench. Mixed-data training improves robustness, but larger mixtures do not dominate across decoding temperatures. We also study how to combine specialized drafters at inference time. Naive checkpoint averaging performs poorly, whereas confidence-based routing improves over single-domain drafts and merged-tree verification yields the highest acceptance length overall for both backbones. Finally, confidence is a more useful routing signal than entropy: rejected tokens tend to have higher entropy, but confidence produces much clearer benchmark-level routing decisions. These results show that speculative decoding quality depends not only on draft architecture, but also on the match between draft training data and downstream workload, and that specialized drafters are better combined at inference time than in weight space.

Diffusion-based vision-language-action models (dVLAs) are promising for embodied intelligence but are fundamentally limited in real-time deployment by the high latency of full inference. We propose Realtime-VLA FLASH, a speculative inference framework that eliminates most full inference calls during replanning by introducing a lightweight draft model with parallel verification via the main model's Action Expert and a phase-aware fallback mechanism that reverts to the full inference pipeline when needed. This design enables low-latency, high-frequency replanning without sacrificing reliability. Experiments show that on LIBERO, FLASH largely preserves task performance by replacing many 58.0 ms full-inference rounds with speculative rounds as fast as 7.8 ms, lowering task-level average inference latency to 19.1 ms (3.04x speedup). We additionally demonstrate effectiveness on real-world conveyor-belt sorting, highlighting its practical impact for latency-critical embodied tasks.

Speculative decoding (SD) has emerged as a promising technique to accelerate LLM inference by employing a small draft model to propose draft tokens in advance, and validating them in parallel with the large target model. However, the existing SD methods still remain constrained by their serialized execution, which causes the mutual waiting bubbles between the draft and target models. To address this challenge, we draw inspiration from branch prediction in modern processors and propose a novel framework SpecBranch to unlock branch parallelism in SD. Specifically, we first take an in-depth analysis of the potential of branch parallelism in SD, and recognize that the key challenge lies in the trade-offs between parallelization and token rollback. Based on the analysis, we introduce parallel speculative branches to preemptively hedge against likely rejections. Meanwhile, to enhance parallelism, we jointly orchestrate adaptive draft lengths with a hybrid combination of the implicit draft model confidence and explicit reusing of target model features. Extensive experiments across various models and benchmarks show that SpecBranch achieves over 1.8times sim 4.5times speedups against the auto-regressive decoding and reduces rollback tokens by 50\% for poorly aligned models, while maintaining an identical sampling distribution.

Deploying large language models (LLMs) in mobile and edge computing environments is constrained by limited on-device resources, scarce wireless bandwidth, and frequent model evolution. Although edge-cloud collaborative inference with speculative decoding (SD) can reduce end-to-end latency by executing a lightweight draft model at the edge and verifying it with a cloud-side target model, existing frameworks fundamentally rely on tight coupling between the two models. Consequently, repeated model synchronization introduces excessive communication overhead, increasing end-to-end latency, and ultimately limiting the scalability of SD in edge environments. To address these limitations, we propose FlexSpec, a communication-efficient collaborative inference framework tailored for evolving edge-cloud systems. The core design of FlexSpec is a shared-backbone architecture that allows a single and static edge-side draft model to remain compatible with a large family of evolving cloud-side target models. By decoupling edge deployment from cloud-side model updates, FlexSpec eliminates the need for edge-side retraining or repeated model downloads, substantially reducing communication and maintenance costs. Furthermore, to accommodate time-varying wireless conditions and heterogeneous device constraints, we develop a channel-aware adaptive speculation mechanism that dynamically adjusts the speculative draft length based on real-time channel state information and device energy budgets. Extensive experiments demonstrate that FlexSpec achieves superior performance compared to conventional SD approaches in terms of inference efficiency.

Large Language Models (LLMs) have revolutionized natural language processing by unifying tasks into text generation, yet their large parameter sizes and autoregressive nature limit inference speed. SAM-Decoding addresses this by introducing a novel retrieval-based speculative decoding method that uses a suffix automaton for efficient and accurate draft generation. Unlike n-gram matching used by the existing method, SAM-Decoding finds the longest suffix match in generating text and text corpuss, achieving an average time complexity of O(1) per generation step. SAM-Decoding constructs static and dynamic suffix automatons for the text corpus and input prompts, respectively, enabling fast and precise draft generation. Meanwhile, it is designed as an approach that can be combined with existing methods, allowing SAM-Decoding to adaptively select a draft generation strategy based on the matching length, thus increasing the inference speed of the LLM. When combined with Token Recycling, evaluations show SAM-Decoding outperforms existing model-free methods, achieving a speedup of 2.27times over autoregressive decoding on Spec-Bench. When combined with EAGLE2, it reaches a speedup of 2.49times, surpassing all current approaches. Our code is available at https://github.com/hyx1999/SAM-Decoding.

Personalized text generation requires models not only to produce coherent text but also to align with a target user's style, tone, and topical focus. Existing retrieval-augmented approaches such as LaMP and PGraphRAG enrich profiles with user and neighbor histories, but they stop at generation and often yield outputs that drift in tone, topic, or style. We present PerFine, a unified, training-free critique-refine framework that enhances personalization through iterative, profile-grounded feedback. In each iteration, an LLM generator produces a draft conditioned on the retrieved profile, and a critic LLM - also conditioned on the same profile - provides structured feedback on tone, vocabulary, sentence structure, and topicality. The generator then revises, while a novel knockout strategy retains the stronger draft across iterations. We further study additional inference-time strategies such as Best-of-N and Topic Extraction to balance quality and efficiency. Across Yelp, Goodreads, and Amazon datasets, PerFine consistently improves personalization over PGraphRAG, with GEval gains of +7-13%, steady improvements over 3-5 refinement iterations, and scalability with increasing critic size. These results highlight that post-hoc, profile-aware feedback offers a powerful paradigm for personalized LLM generation that is both training-free and model-agnostic.

We introduce ``Idea to Image,'' a system that enables multimodal iterative self-refinement with GPT-4V(ision) for automatic image design and generation. Humans can quickly identify the characteristics of different text-to-image (T2I) models via iterative explorations. This enables them to efficiently convert their high-level generation ideas into effective T2I prompts that can produce good images. We investigate if systems based on large multimodal models (LMMs) can develop analogous multimodal self-refinement abilities that enable exploring unknown models or environments via self-refining tries. Idea2Img cyclically generates revised T2I prompts to synthesize draft images, and provides directional feedback for prompt revision, both conditioned on its memory of the probed T2I model's characteristics. The iterative self-refinement brings Idea2Img various advantages over vanilla T2I models. Notably, Idea2Img can process input ideas with interleaved image-text sequences, follow ideas with design instructions, and generate images of better semantic and visual qualities. The user preference study validates the efficacy of multimodal iterative self-refinement on automatic image design and generation.

We introduce ReflectDrive-2, a masked discrete diffusion planner with separate action expert for autonomous driving that represents plans as discrete trajectory tokens and generates them through parallel masked decoding. This discrete token space enables in-place trajectory revision: AutoEdit rewrites selected tokens using the same model, without requiring an auxiliary refinement network. To train this capability, we use a two-stage procedure. First, we construct structure-aware perturbations of expert trajectories along longitudinal progress and lateral heading directions and supervise the model to recover the original expert trajectory. We then fine-tune the full decision--draft--reflect rollout with reinforcement learning (RL), assigning terminal driving reward to the final post-edit trajectory and propagating policy-gradient credit through full-rollout transitions. Full-rollout RL proves crucial for coupling drafting and editing: under supervised training alone, inference-time AutoEdit improves PDMS by at most 0.3, whereas RL increases its gain to 1.9. We also co-design an efficient reflective decoding stack for the decision--draft--reflect pipeline, combining shared-prefix KV reuse, Alternating Step Decode, and fused on-device unmasking. On NAVSIM, ReflectDrive-2 achieves 91.0 PDMS with camera-only input and 94.8 PDMS in a best-of-6 oracle setting, while running at 31.8 ms average latency on NVIDIA Thor.

We release Super Apriel, a 15B-parameter supernet in which every decoder layer provides four trained mixer choices -- Full Attention (FA), Sliding Window Attention (SWA), Kimi Delta Attention (KDA), and Gated DeltaNet (GDN). A placement selects one mixer per layer; placements can be switched between requests at serving time without reloading weights, enabling multiple speed presets from a single checkpoint. The shared checkpoint also enables speculative decoding without a separate draft model. The all-FA preset matches the Apriel 1.6 teacher on all reported benchmarks; recommended hybrid presets span 2.9times to 10.7times decode throughput at 96% to 77% quality retention, with throughput advantages that compound at longer context lengths. With four mixer types across 48 layers, the configuration space is vast. A surrogate that predicts placement quality from the per-layer mixer assignment makes the speed-quality landscape tractable and identifies the best tradeoffs at each speed level. We investigate whether the best configurations at each speed level can be identified early in training or only after convergence. Rankings stabilize quickly at 0.5B scale, but the most efficient configurations exhibit higher instability at 15B, cautioning against extrapolation from smaller models. Super Apriel is trained by stochastic distillation from a frozen Apriel 1.6 teacher, followed by supervised fine-tuning. We release the supernet weights, Fast-LLM training code, vLLM serving code, and a placement optimization toolkit.

While autoregressive (AR) modeling has recently emerged as a new paradigm in visual generation, its practical adoption is severely constrained by the slow inference speed of per-token generation, which often requires thousands of steps to produce a single sample. To address this challenge, we propose MC-SJD, a training-free, lossless parallel decoding framework designed to accelerate AR visual generation by extending the recently introduced Speculative Jacobi Decoding (SJD). Although SJD shows strong potential for accelerating AR generation, we demonstrate that token instability across iterations significantly reduces the acceptance rate, a limitation that primarily arises from the independent sampling process used during draft token generation. To overcome this, we introduce MC-SJD, an information-theoretic approach based on coupling, which substantially accelerates standard SJD by maximizing the probability of sampling identical draft tokens across consecutive iterations, all while preserving its lossless property. Remarkably, this method requires only a single-line modification to the existing algorithm, yet achieves substantial performance gains, delivering up to a ~4.2x acceleration in image generation and ~13.3x acceleration in video generation compared to standard AR decoding, without any degradation in output quality.

Parallel Speculative Decoding (PSD) accelerates traditional Speculative Decoding (SD) by overlapping draft generation with verification. However, it remains hampered by two fundamental challenges: (1) a theoretical speedup ceiling dictated by the speed ratio between the draft and target models, and (2) high computational waste and pipeline stall due to mid-sequence token rejections of early errors. To address these limitations, we introduce Double (Double Retrieval Speculative Parallelism). By bridging the gap between SD and PSD, our framework resolves the Retrieval Precision-Efficiency Dilemma through a novel synchronous mechanism. Specifically, we enable the draft model to execute iterative retrieval speculations to break the theoretical speedup limits; to alleviate rejections without rollback, the target model performs authoritative retrieval to generate multi-token guidance. Double is entirely training-free and lossless. Extensive experiments demonstrate state-of-the-art speedup of 5.3times on LLaMA3.3-70B and 2.8times on Qwen3-32B, significantly outperforming the advanced method EAGLE-3 that requires extensive model training.

Speculative decoding is a relatively new decoding framework that leverages small and efficient draft models to reduce the latency of LLMs. In this study, we introduce GliDe and CaPE, two low-hassle modifications to vanilla speculative decoding to further improve the decoding speed of a frozen LLM. Specifically, GliDe is a modified draft model architecture that reuses the cached keys and values from the target LLM, while CaPE is a proposal expansion method that uses the draft model's confidence scores to help select additional candidate tokens for verification. Extensive experiments on different benchmarks demonstrate that our proposed GliDe draft model significantly reduces the expected decoding latency. Additional evaluation using walltime reveals that GliDe can accelerate Vicuna models up to 2.17x and further extend the improvement to 2.61x with CaPE. We will release our code, data, and the trained draft models.

Diffusion large language models (dLLMs) have recently emerged as a promising alternative to autoregressive (AR) LLMs, offering faster inference through parallel or blockwise decoding. However, their masked language modeling formulation remains incompatible with standard token-level speculative decoding, one of the most effective acceleration techniques for AR models. In AR decoding, the causal mask preserves temporally valid token-level contexts, enabling a target model to verify multiple drafted tokens in a single forward pass. In contrast, dLLMs rely on mask tokens and bidirectional attention, causing the effective context to change across denoising steps and preventing direct token-level speculative verification. To bridge this gap, we propose a simple but effective speculative decoding algorithm for diffusion language models, named SimSD, which mainly adopts a plug-and-play masking strategy that equips dLLMs with temporally valid token-level contexts for speculative decoding. Our method explicitly introduces reference tokens from draft-model predictions and designs an attention mask that regulates their interaction with current-step tokens, allowing dLLMs to compute valid logits for drafted tokens in a single forward pass. This restores the key verification ability provided by causal masking in AR models while preserving the parallel decoding advantages of dLLMs. The proposed method is training-free and can be flexibly integrated with other acceleration techniques such as KV cache and blockwise decoding. Experiments on SDAR-family dLLMs across four benchmarks show that our method achieves up to 7.46x higher decoding throughput while maintaining and even improving average generation quality.

The immense model sizes of large language models (LLMs) challenge deployment on memory-limited consumer GPUs. Although model compression and parameter offloading are common strategies to address memory limitations, compression can degrade quality, and offloading maintains quality but suffers from slow inference. Speculative decoding presents a promising avenue to accelerate parameter offloading, utilizing a fast draft model to propose multiple draft tokens, which are then verified by the target LLM in parallel with a single forward pass. This method reduces the time-consuming data transfers in forward passes that involve offloaded weight transfers. Existing methods often rely on pretrained weights of the same family, but require additional training to align with custom-trained models. Moreover, approaches that involve draft model training usually yield only modest speedups. This limitation arises from insufficient alignment with the target model, preventing higher token acceptance lengths. To address these challenges and achieve greater speedups, we propose SubSpec, a plug-and-play method to accelerate parameter offloading that is lossless and training-free. SubSpec constructs a highly aligned draft model by generating low-bit quantized substitute layers from offloaded target LLM portions. Additionally, our method shares the remaining GPU-resident layers and the KV-Cache, further reducing memory overhead and enhance alignment. SubSpec achieves a high average acceptance length, delivering 9.1x speedup for Qwen2.5 7B on MT-Bench (8GB VRAM limit) and an average of 12.5x speedup for Qwen2.5 32B on popular generation benchmarks (24GB VRAM limit).

To support emerging language-based applications using dispersed and heterogeneous computing resources, the hybrid language model (HLM) offers a promising architecture, where an on-device small language model (SLM) generates draft tokens that are validated and corrected by a remote large language model (LLM). However, the original HLM suffers from substantial communication overhead, as the LLM requires the SLM to upload the full vocabulary distribution for each token. Moreover, both communication and computation resources are wasted when the LLM validates tokens that are highly likely to be accepted. To overcome these limitations, we propose communication-efficient and uncertainty-aware HLM (CU-HLM). In CU-HLM, the SLM transmits truncated vocabulary distributions only when its output uncertainty is high. We validate the feasibility of this opportunistic transmission by discovering a strong correlation between SLM's uncertainty and LLM's rejection probability. Furthermore, we theoretically derive optimal uncertainty thresholds and optimal vocabulary truncation strategies. Simulation results show that, compared to standard HLM, CU-HLM achieves up to 206times higher token throughput by skipping 74.8% transmissions with 97.4% vocabulary compression, while maintaining 97.4% accuracy.

Recently, large language models (LLMs) have demonstrated remarkable performance in abstractive summarization tasks. However, controllable summarization with LLMs remains underexplored, limiting their ability to generate summaries that align with specific user preferences. In this paper, we first investigate the capability of LLMs to control diverse attributes, revealing that they encounter greater challenges with numerical attributes, such as length and extractiveness, compared to linguistic attributes. To address this challenge, we propose a guide-to-explain framework (GTE) for controllable summarization. Our GTE framework enables the model to identify misaligned attributes in the initial draft and guides it in explaining errors in the previous output. Based on this reflection, the model generates a well-adjusted summary. As a result, by allowing the model to reflect on its misalignment, we generate summaries that satisfy the desired attributes in surprisingly fewer iterations than other iterative methods solely using LLMs.

While large language models (LLMs) like GPT-4 have recently demonstrated astonishing zero-shot capabilities in general domain tasks, they often generate content with hallucinations in specific domains such as Chinese law, hindering their application in these areas. This is typically due to the absence of training data that encompasses such a specific domain, preventing GPT-4 from acquiring in-domain knowledge. A pressing challenge is that it's not plausible to continue training LLMs of such scale on in-domain data. This paper introduces a simple and effective domain adaptation framework for GPT-4 by reformulating generation as an adapt-retrieve-revise process. The initial step is to adapt an affordable 7B LLM to the target domain by continuing learning on in-domain data. When solving a task, we leverage the adapted LLM to generate a draft answer given a task query. Then, the draft answer will be used to retrieve supporting evidence candidates from an external in-domain knowledge base. Finally, the draft answer and retrieved evidence are concatenated into a whole prompt to let GPT-4 assess the evidence and revise the draft answer to generate the final answer. Our proposal combines the advantages of the efficiency of adapting a smaller 7B model with the evidence-assessing capability of GPT-4 and effectively prevents GPT-4 from generating hallucinatory content. In the zero-shot setting of four Chinese legal tasks, our method improves accuracy by 33.3\% compared to the direct generation by GPT-4. When compared to two stronger retrieval-based baselines, our method outperforms them by 15.4\% and 23.9\%. Our code will be released

Documentation burden is a major contributor to clinician burnout, which is rising nationally and is an urgent threat to our ability to care for patients. Artificial intelligence (AI) chatbots, such as ChatGPT, could reduce clinician burden by assisting with documentation. Although many hospitals are actively integrating such systems into electronic medical record systems, AI chatbots utility and impact on clinical decision-making have not been studied for this intended use. We are the first to examine the utility of large language models in assisting clinicians draft responses to patient questions. In our two-stage cross-sectional study, 6 oncologists responded to 100 realistic synthetic cancer patient scenarios and portal messages developed to reflect common medical situations, first manually, then with AI assistance. We find AI-assisted responses were longer, less readable, but provided acceptable drafts without edits 58% of time. AI assistance improved efficiency 77% of time, with low harm risk (82% safe). However, 7.7% unedited AI responses could severely harm. In 31% cases, physicians thought AI drafts were human-written. AI assistance led to more patient education recommendations, fewer clinical actions than manual responses. Results show promise for AI to improve clinician efficiency and patient care through assisting documentation, if used judiciously. Monitoring model outputs and human-AI interaction remains crucial for safe implementation.

Large Language Models (LLMs) have shown promise in solving complex mathematical problems, yet they still fall short of producing accurate and consistent solutions. Reinforcement Learning (RL) is a framework for aligning these models with task-specific rewards, improving overall quality and reliability. Group Relative Policy Optimization (GRPO) is an efficient, value-function-free alternative to Proximal Policy Optimization (PPO) that leverages group-relative reward normalization. We introduce Iterative Group Relative Policy Optimization (iGRPO), a two-stage extension of GRPO that adds dynamic self-conditioning through model-generated drafts. In Stage 1, iGRPO samples multiple exploratory drafts and selects the highest-reward draft using the same scalar reward signal used for optimization. In Stage 2, it appends this best draft to the original prompt and applies a GRPO-style update on draft-conditioned refinements, training the policy to improve beyond its strongest prior attempt. Under matched rollout budgets, iGRPO consistently outperforms GRPO across base models (e.g., Nemotron-H-8B-Base-8K and DeepSeek-R1 Distilled), validating its effectiveness on diverse reasoning benchmarks. Moreover, applying iGRPO to OpenReasoning-Nemotron-7B trained on AceReason-Math achieves new state-of-the-art results of 85.62\% and 79.64\% on AIME24 and AIME25, respectively. Ablations further show that the refinement wrapper generalizes beyond GRPO variants, benefits from a generative judge, and alters learning dynamics by delaying entropy collapse. These results underscore the potential of iterative, self-feedback-based RL for advancing verifiable mathematical reasoning.

Lane-level maps are critical infrastructure for autonomous driving and lane-level navigation, yet constructing and maintaining standardized lane networks for hundreds of cities remains highly labor-intensive. Recent end-to-end vectorized mapping methods can predict lane geometry and topology directly from sensor data, but they typically treat mapping specifications and traffic regulations as implicit, dataset-dependent supervision. Moreover, in complex scenes (e.g., worn or missing markings and occlusions), correct lane configurations are often under-determined by visual evidence alone, making specification violations a major source of human post-editing. We propose MapAgent, an industrial-grade agentic architecture that augments a vectorization backbone for specification-compliant lane-map production. Rather than merely adding an agent loop to map prediction, MapAgent couples backbone perception with explicit specification verification, constraint-aware reasoning, and deterministic map editing under a bounded, verification-driven Judge-Planner-Worker loop. A vision-language Judge diagnoses errors by jointly inspecting visual evidence and draft vectors, while a tool-calling Planner generates minimal corrective edits with post-edit re-validation. To remain scalable for city-scale production, MapAgent is selectively triggered only on tiles with low backbone confidence, adding modest overhead while preserving throughput. Experiments on real-world datasets show consistent gains over strong production baselines, especially in complex and long-tail scenarios. Additionally, MapAgent has been integrated into Baidu Maps, supporting lane-level map generation for over 360 cities nationwide and elevating the overall production automation to over 95%, demonstrating MapAgent's practicality and effectiveness for large-scale lane-level map generation.

Speculative decoding enhances the efficiency of large language models (LLMs) by leveraging a draft model to draft for a larger target model to review. However, drafting in speculative decoding involves slow autoregressive generation and generating tokens of different importance with the same time allocation. These two inefficiencies lead to its suboptimal performance. To address this issue, we introduce Cascade Speculative Drafting (CS. Drafting), a novel approach that employs two types of cascades. The Vertical Cascade eliminates autoregressive generation from neural models. The Horizontal Cascade constitutes efficient time allocation in drafting with its optimality supported by our theoretical analysis. Combining both cascades, our CS. Drafting algorithm has achieved up to 72 percent additional speedup over speculative decoding in our experiments while keeping the same output distribution.

Speculative decoding is a promising method for reducing the inference latency of large language models. The effectiveness of the method depends on the speculation length (SL) - the number of tokens generated by the draft model at each iteration. The vast majority of speculative decoding approaches use the same SL for all iterations. In this work, we show that this practice is suboptimal. We introduce DISCO, a DynamIc SpeCulation length Optimization method that uses a classifier to dynamically adjust the SL at each iteration, while provably preserving the decoding quality. Experiments with four benchmarks demonstrate average speedup gains of 10.3% relative to our best baselines.

Natural language to repository generation (NL2Repo) requires a system to construct an entire software repository from a natural-language requirements document. Compared with function-level code generation, this task demands longer planning horizons, stable interfaces across files, and iterative debugging of cross-file inconsistencies. To address these challenges, we propose CodeTeam, an LLM-based multi-agent framework that separates planning, decision making, and implementation into distinct, coordinated stages. In the planning stage, multiple Architect agents draft competing software design sketches (SDS), optionally grounded by retrieved design references. A CTO agent then evaluates, selects, and normalizes the most promising SDS into a machine-checkable contract that specifies file ownership, public interfaces, and dependency constraints. In the implementation stage, Developer agents generate code under a dependency-aware scheduler with bounded context and lightweight Git-based coordination, while a QA agent runs tests and drives iterative repairs. On the synthesis-based SketchEval benchmark, we explicitly compare CodeTeam's prompt-engineering (PE) and supervised fine-tuning (SFT) variants with the corresponding CodeS variants, where CodeTeam improves the overall SketchBLEU by 4.1 and 2.9 absolute points, respectively. On the execution-based NL2Repo-Bench benchmark, used as an external validation protocol, CodeTeam achieves the highest average test pass rate in both settings (34.6% PE, 42.3% SFT), confirming that the sketch-improvements extend to functional correctness under upstream test suites. Ablation results show that project-specific developer allocation and retrieval-augmented planning each contribute substantially to the SketchBLEU improvement (9.9% and 8.1% relative, respectively). CodeTeam and the experimental results are available at https://github.com/WhitenWhiten/CodeTeam

Speculative decoding accelerates large language model inference by proposing tokens with a lightweight draft model and selectively accepting them using a target model. This work introduces DropMatch, a novel approach that matches draft tokens to the predictive distribution of the target model via Monte Carlo dropout applied exclusively to the LM head, enabling sampling-based acceptance decisions. By generating multiple decoding paths, our method forms an empirical token distribution against which draft tokens are evaluated for consistency. This acceptance mechanism enables the model to adaptively control the size of decoding paths under an appropriate dropout probability, preventing substantial distortion of the target model predictive distribution. The proposed method operates in a training-free, data-free, and calibration-free manner, requires no architectural modification to pretrained models, and can be orthogonally integrated with a wide range of existing speculative decoding and inference acceleration techniques. Experiments across multiple benchmarks demonstrate that our approach increases acceptance length while maintaining competitive task performance, yielding inference speedups ranging from 1.09x to 1.33x over the standard baseline, and up to an additional 1.09x speedup when applied on top of EAGLE3.

We propose a novel speculative decoding method tailored for multi-sample reasoning scenarios, such as self-consistency and Best-of-N sampling. Our method exploits the intrinsic consensus of parallel generation paths to synthesize high-quality draft tokens without requiring auxiliary models or external databases. By dynamically analyzing structural patterns across parallel reasoning paths through a probabilistic aggregation mechanism, it identifies consensus token sequences that align with the decoding distribution. Evaluations on mathematical reasoning benchmarks demonstrate a substantial improvement in draft acceptance rates over baselines, while reducing the latency in draft token construction. This work establishes a paradigm shift for efficient multi-sample inference, enabling seamless integration of speculative decoding with sampling-based reasoning techniques.

Humans paint images incrementally: they plan a global layout, sketch a coarse draft, inspect, and refine details, and most importantly, each step is grounded in the evolving visual states. However, can unified multimodal models trained on text-image interleaved datasets also imagine the chain of intermediate states? In this paper, we introduce process-driven image generation, a multi-step paradigm that decomposes synthesis into an interleaved reasoning trajectory of thoughts and actions. Rather than generating images in a single step, our approach unfolds across multiple iterations, each consisting of 4 stages: textual planning, visual drafting, textual reflection, and visual refinement. The textual reasoning explicitly conditions how the visual state should evolve, while the generated visual intermediate in turn constrains and grounds the next round of textual reasoning. A core challenge of process-driven generation stems from the ambiguity of intermediate states: how can models evaluate each partially-complete image? We address this through dense, step-wise supervision that maintains two complementary constraints: for the visual intermediate states, we enforce the spatial and semantic consistency; for the textual intermediate states, we preserve the prior visual knowledge while enabling the model to identify and correct prompt-violating elements. This makes the generation process explicit, interpretable, and directly supervisable. To validate proposed method, we conduct experiments under various text-to-image generation benchmarks.

We present MiMo-V2-Flash, a Mixture-of-Experts (MoE) model with 309B total parameters and 15B active parameters, designed for fast, strong reasoning and agentic capabilities. MiMo-V2-Flash adopts a hybrid attention architecture that interleaves Sliding Window Attention (SWA) with global attention, with a 128-token sliding window under a 5:1 hybrid ratio. The model is pre-trained on 27 trillion tokens with Multi-Token Prediction (MTP), employing a native 32k context length and subsequently extended to 256k. To efficiently scale post-training compute, MiMo-V2-Flash introduces a novel Multi-Teacher On-Policy Distillation (MOPD) paradigm. In this framework, domain-specialized teachers (e.g., trained via large-scale reinforcement learning) provide dense and token-level reward, enabling the student model to perfectly master teacher expertise. MiMo-V2-Flash rivals top-tier open-weight models such as DeepSeek-V3.2 and Kimi-K2, despite using only 1/2 and 1/3 of their total parameters, respectively. During inference, by repurposing MTP as a draft model for speculative decoding, MiMo-V2-Flash achieves up to 3.6 acceptance length and 2.6x decoding speedup with three MTP layers. We open-source both the model weights and the three-layer MTP weights to foster open research and community collaboration.

Modern Large Language Models achieve impressive reasoning capabilities with long Chain of Thoughts, but they incur substantial computational cost during inference, and this motivates techniques to improve the performance-cost ratio. Among these techniques, Speculative Decoding accelerates inference by employing a fast but inaccurate draft model to autoregressively propose tokens, which are then verified in parallel by a more capable target model. However, due to unnecessary rejections caused by token mismatches in semantically equivalent steps, traditional token-level Speculative Decoding struggles in reasoning tasks. Although recent works have shifted to step-level semantic verification, which improve efficiency by accepting or rejecting entire reasoning steps, existing step-level methods still regenerate many rejected steps with little improvement, wasting valuable target compute. To address this challenge, we propose Arbitrage, a novel step-level speculative generation framework that routes generation dynamically based on the relative advantage between draft and target models. Instead of applying a fixed acceptance threshold, Arbitrage uses a lightweight router trained to predict when the target model is likely to produce a meaningfully better step. This routing approximates an ideal Arbitrage Oracle that always chooses the higher-quality step, achieving near-optimal efficiency-accuracy trade-offs. Across multiple mathematical reasoning benchmarks, Arbitrage consistently surpasses prior step-level Speculative Decoding baselines, reducing inference latency by up to sim2times at matched accuracy.

Speculative decoding has been widely used to accelerate autoregressive (AR) text generation. However, its effectiveness in visual AR models remains limited due to token selection ambiguity, where multiple tokens receive similarly low probabilities, reducing acceptance rates. While dynamic tree drafting has been proposed to improve speculative decoding, we show that it fails to mitigate token selection ambiguity, resulting in shallow draft trees and suboptimal acceleration. To address this, we introduce LANTERN++, a novel framework that integrates static tree drafting with a relaxed acceptance condition, allowing drafts to be selected independently of low-confidence predictions. This enables deeper accepted sequences, improving decoding efficiency while preserving image quality. Extensive experiments on state-of-the-art visual AR models demonstrate that LANTERN++ significantly accelerates inference, achieving up to times 2.56 speedup over standard AR decoding while maintaining high image quality.

The performance of large language models (LLMs) is closely linked to their underlying size, leading to ever-growing networks and hence slower inference. Speculative decoding has been proposed as a technique to accelerate autoregressive generation, leveraging a fast draft model to propose candidate tokens, which are then verified in parallel based on their likelihood under the target model. While this approach guarantees to reproduce the target output, it incurs a substantial penalty: many high-quality draft tokens are rejected, even when they represent objectively valid continuations. Indeed, we show that even powerful draft models such as GPT-4o, as well as human text cannot achieve high acceptance rates under the standard verification scheme. This severely limits the speedup potential of current speculative decoding methods, as an early rejection becomes overwhelmingly likely when solely relying on alignment of draft and target. We thus ask the following question: Can we adapt verification to recognize correct, but non-aligned replies? To this end, we draw inspiration from the LLM-as-a-judge framework, which demonstrated that LLMs are able to rate answers in a versatile way. We carefully design a dataset to elicit the same capability in the target model by training a compact module on top of the embeddings to produce ``judgements" of the current continuation. We showcase our strategy on the Llama-3.1 family, where our 8b/405B-Judge achieves a speedup of 9x over Llama-405B, while maintaining its quality on a large range of benchmarks. These benefits remain present even in optimized inference frameworks, where our method reaches up to 141 tokens/s for 8B/70B-Judge and 129 tokens/s for 8B/405B on 2 and 8 H100s respectively.

Accelerating inference in Large Language Models (LLMs) is critical for real-time interactions, as they have been widely incorporated into real-world services. Speculative decoding, a fully algorithmic solution, has gained attention for improving inference speed by drafting and verifying tokens, thereby generating multiple tokens in a single forward pass. However, current drafting strategies usually require significant fine-tuning or have inconsistent performance across tasks. To address these challenges, we propose Hierarchy Drafting (HD), a novel lossless drafting approach that organizes various token sources into multiple databases in a hierarchical framework based on temporal locality. In the drafting step, HD sequentially accesses multiple databases to obtain draft tokens from the highest to the lowest locality, ensuring consistent acceleration across diverse tasks and minimizing drafting latency. Our experiments on Spec-Bench using LLMs with 7B and 13B parameters demonstrate that HD outperforms existing database drafting methods, achieving robust inference speedups across model sizes, tasks, and temperatures.

Despite their strong performance on reasoning tasks, large reasoning models (LRMs) often suffer from overthinking, producing unnecessarily long outputs and incurring high end-to-end latency, a significant limitation to their real-world deployment. To address overthinking, early-exit mechanisms have been proposed to terminate reasoning before typical completion, showing that this approach can effectively shorten generation length with minimal impact on accuracy. However, their reliance on probing mechanisms introduces a detection overhead that limits their end-to-end latency gains and compromises their generalizability across diverse problems. Inspired by the use of hidden states in speculative decoding, we propose SpecExit, a novel framework that predicts both future tokens and an early-exit signal directly from a lightweight draft model without probing overhead. Our method offers significant improvements, reducing average generation length by 66\% and achieving a 2.5x speedup in end-to-end latency compared to the speculative decoding baseline, without compromising accuracy. Our method leverages the inherent signals from hidden states to provide effective early-exit signals, suggesting broader use of hidden states for efficient reasoning. Our code is available at https://github.com/Tencent/AngelSlim.

Reducing the inference latency of large language models (LLMs) is crucial, and speculative decoding (SD) stands out as one of the most effective techniques. Rather than letting the LLM generate all tokens directly, speculative decoding employs effective proxies to predict potential outputs, which are then verified by the LLM without compromising the generation quality. Yet, deploying SD in real online LLM serving systems (with continuous batching) does not always yield improvement -- under higher request rates or low speculation accuracy, it paradoxically increases latency. Furthermore, there is no best speculation length work for all workloads under different system loads. Based on the observations, we develop a dynamic framework SmartSpec. SmartSpec dynamically determines the best speculation length for each request (from 0, i.e., no speculation, to many tokens) -- hence the associated speculative execution costs -- based on a new metric called goodput, which characterizes the current observed load of the entire system and the speculation accuracy. We show that SmartSpec consistently reduces average request latency by up to 3.2x compared to non-speculative decoding baselines across different sizes of target models, draft models, request rates, and datasets. Moreover, SmartSpec can be applied to different styles of speculative decoding, including traditional, model-based approaches as well as model-free methods like prompt lookup and tree-style decoding.

Large Language Models (LLMs) have achieved significant advances in natural language processing, yet their potential for high-stake political decision-making remains largely unexplored. This paper addresses the gap by focusing on the application of LLMs to the United Nations (UN) decision-making process, where the stakes are particularly high and political decisions can have far-reaching consequences. We introduce a novel dataset comprising publicly available UN Security Council (UNSC) records from 1994 to 2024, including draft resolutions, voting records, and diplomatic speeches. Using this dataset, we propose the United Nations Benchmark (UNBench), the first comprehensive benchmark designed to evaluate LLMs across four interconnected political science tasks: co-penholder judgment, representative voting simulation, draft adoption prediction, and representative statement generation. These tasks span the three stages of the UN decision-making process--drafting, voting, and discussing--and aim to assess LLMs' ability to understand and simulate political dynamics. Our experimental analysis demonstrates the potential and challenges of applying LLMs in this domain, providing insights into their strengths and limitations in political science. This work contributes to the growing intersection of AI and political science, opening new avenues for research and practical applications in global governance. The UNBench Repository can be accessed at: https://github.com/yueqingliang1/UNBench.

Large reasoning models (LRMs) achieve strong reasoning performance by emitting long chains of thought. Yet, these verbose traces slow down inference and often drift into unnecessary detail, known as the overthinking phenomenon. To better understand LRMs' behavior, we systematically analyze the token-level misalignment between reasoning and non-reasoning models. While it is expected that their primary difference lies in the stylistic "thinking cues", LRMs uniquely exhibit two pivotal, previously under-explored phenomena: a Global Misalignment Rebound, where their divergence from non-reasoning models persists or even grows as response length increases, and more critically, a Local Misalignment Diminish, where the misalignment concentrates at the "thinking cues" each sentence starts with but rapidly declines in the remaining of the sentence. Motivated by the Local Misalignment Diminish, we propose FoReaL-Decoding, a collaborative fast-slow thinking decoding method for cost-quality trade-off. In FoReaL-Decoding, a Leading model leads the first few tokens for each sentence, and then a weaker draft model completes the following tokens to the end of each sentence. FoReaL-Decoding adopts a stochastic gate to smoothly interpolate between the small and the large model. On four popular math-reasoning benchmarks (AIME24, GPQA-Diamond, MATH500, AMC23), FoReaL-Decoding reduces theoretical FLOPs by 30 to 50% and trims CoT length by up to 40%, while preserving 86 to 100% of model performance. These results establish FoReaL-Decoding as a simple, plug-and-play route to controllable cost-quality trade-offs in reasoning-centric tasks.

Information retrieval, specifically contract clause retrieval, is foundational to contract drafting because lawyers rarely draft contracts from scratch; instead, they locate and revise the most relevant precedent. We introduce the Atticus Clause Retrieval Dataset (ACORD), the first retrieval benchmark for contract drafting fully annotated by experts. ACORD focuses on complex contract clauses such as Limitation of Liability, Indemnification, Change of Control, and Most Favored Nation. It includes 114 queries and over 126,000 query-clause pairs, each ranked on a scale from 1 to 5 stars. The task is to find the most relevant precedent clauses to a query. The bi-encoder retriever paired with pointwise LLMs re-rankers shows promising results. However, substantial improvements are still needed to effectively manage the complex legal work typically undertaken by lawyers. As the first retrieval benchmark for contract drafting annotated by experts, ACORD can serve as a valuable IR benchmark for the NLP community.