Daily Papers

Knowledge distillation (KD) is a standard route to compress Large Language Models (LLMs) into compact students, yet most pipelines uniformly apply token-wise loss regardless of teacher confidence. This indiscriminate supervision amplifies noisy, high-entropy signals and is especially harmful under large teacher-student capacity gaps. We introduce SelecTKD, a plug-and-play Selective Token-Weighted distillation framework that shifts the focus from "how to measure divergence" to "where to apply learning". At each step, the student proposes tokens that are verified by the teacher through a robust propose-and-verify procedure with two variants: greedy Top-k and non-greedy Spec-k. Accepted tokens receive full loss, while rejected tokens are masked or down-weighted. This objective-agnostic design works with on- and off-policy data, induces an implicit curriculum quantified by Token Acceptance Rate (TAR), and stabilizes optimization. Across instruction following, mathematical reasoning, code generation, and a VLM setting, SelecTKD consistently improves strong baselines and achieves state-of-the-art results for small models without architectural changes or extra reference models.

While speculative decoding has recently appeared as a promising direction for accelerating the inference of large language models (LLMs), the speedup and scalability are strongly bounded by the token acceptance rate. Prevalent methods usually organize predicted tokens as independent chains or fixed token trees, which fails to generalize to diverse query distributions. In this paper, we propose DySpec, a faster speculative decoding algorithm with a novel dynamic token tree structure. We begin by bridging the draft distribution and acceptance rate from intuitive and empirical clues, and successfully show that the two variables are strongly correlated. Based on this, we employ a greedy strategy to dynamically expand the token tree at run time. Theoretically, we show that our method can achieve optimal results under mild assumptions. Empirically, DySpec yields a higher acceptance rate and speedup than fixed trees. DySpec can drastically improve the throughput and reduce the latency of token generation across various data distribution and model sizes, which significantly outperforms strong competitors, including Specinfer and Sequoia. Under low temperature setting, DySpec can improve the throughput up to 9.1times and reduce the latency up to 9.4times on Llama2-70B. Under high temperature setting, DySpec can also improve the throughput up to 6.21times, despite the increasing difficulty of speculating more than one token per step for draft model.

Speculative decoding accelerates Large Language Model (LLM) inference by using a small draft model to predict multiple tokens, and a large target model to verify these tokens in parallel. Recent studies leverage the hidden state of the target model to enhance draft model prediction accuracy. However, existing methods suffer from the degrading quality of draft token predictions at later positions, due to error accumulation in draft model generated features. In this paper, we propose Position Specialists (PosS), which consist of multiple position-specialized draft layers to generate tokens at assigned position(s). Position specialists greatly improve token acceptance rate at later positions per drafting round, as each specialist only needs to focus on handling a certain level of draft model feature deviation. Experiment results on Llama-3-8B-Instruct and Llama-2-13B-chat across six datasets demonstrate that PosS effectively improves over baselines on average acceptance length and speed-up ratio. Our codebase is available at https://github.com/shrango/PosS.

Speculative Decoding (SD) is a widely used approach to accelerate the inference of large language models (LLMs) without reducing generation quality. It operates by first using a compact model to draft multiple tokens efficiently, followed by parallel verification using the target LLM. This approach leads to faster inference compared to auto-regressive decoding. While there are multiple approaches to create a draft model, one promising approach is to use early-exit methods. These methods draft candidate tokens by using a subset of layers of the primary model and applying the remaining layers for verification, allowing a single model to handle both drafting and verification. While this technique reduces memory usage and computational cost, its performance relies on the choice of the exit layer for drafting and the number of tokens drafted (speculation length) in each SD round. Prior works use hyperparameter exploration to statically select these values. However, our evaluations show that these hyperparameter values are task-specific, and even within a task they are dependent on the current sequence context. We introduce DEL, a plug-and-play method that adaptively selects the exit layer and speculation length during inference. DEL dynamically tracks the token acceptance rate if the tokens are drafted at each layer of an LLM and uses that knowledge to heuristically select the optimal exit layer and speculation length. Our experiments across a broad range of models and downstream tasks show that DEL achieves overall speedups of 2.16timessim2.50times over vanilla auto-regressive decoding and improves upon the state-of-the-art SD methods by up to 0.27times.

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 emerged as a powerful approach to accelerate large language model (LLM) inference by employing lightweight draft models to propose candidate tokens that are subsequently verified by the target model. The effectiveness of this paradigm critically depends on the quality of the draft model. While recent advances such as the EAGLE series achieve state-of-the-art speedup, existing draft models remain limited by error accumulation: they condition only on the current prefix, causing their predictions to drift from the target model over steps. In this work, we propose ConFu (Contemplate the Future), a novel speculative decoding framework that enables draft models to anticipate the future direction of generation. ConFu introduces (i) contemplate tokens and soft prompts that allow the draft model to leverage future-oriented signals from the target model at negligible cost, (ii) a dynamic contemplate token mechanism with MoE to enable context-aware future prediction, and (iii) a training framework with anchor token sampling and future prediction replication that learns robust future prediction. Experiments demonstrate that ConFu improves token acceptance rates and generation speed over EAGLE-3 by 8--11% across various downstream tasks with Llama-3 3B and 8B models. We believe our work is the first to bridge speculative decoding with continuous reasoning tokens, offering a new direction for accelerating LLM inference.

Speculative decoding is a pivotal technique to accelerate the inference of large language models (LLMs) by employing a smaller draft model to predict the target model's outputs. However, its efficacy can be limited due to the low predictive accuracy of the draft model, particularly when faced with diverse text inputs and a significant capability gap between the draft and target models. We introduce online speculative decoding (OSD) to address this challenge. The main idea is to continually update (multiple) draft model(s) on observed user query data using the abundant excess computational power in an LLM serving cluster. Given that LLM inference is memory-bounded, the surplus computational power in a typical LLM serving cluster can be repurposed for online retraining of draft models, thereby making the training cost-neutral. Since the query distribution of an LLM service is relatively simple, retraining on query distribution enables the draft model to more accurately predict the target model's outputs, particularly on data originating from query distributions. As the draft model evolves online, it aligns with the query distribution in real time, mitigating distribution shifts. We develop a prototype of online speculative decoding based on online knowledge distillation and evaluate it using both synthetic and real query data on several popular LLMs. The results show a substantial increase in the token acceptance rate by 0.1 to 0.65, which translates into 1.22x to 3.06x latency reduction.

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.

Verification is a key bottleneck in improving inference speed while maintaining distribution fidelity in Speculative Decoding. Recent work has shown that sequence-level verification leads to a higher number of accepted tokens compared to token-wise verification. However, existing solutions often rely on surrogate approximations or are constrained by partial information, struggling with joint intractability. In this work, we propose Hierarchical Speculative Decoding (HSD), a provably lossless verification method that significantly boosts the expected number of accepted tokens and overcomes joint intractability by balancing excess and deficient probability mass across accessible branches. Our extensive large-scale experiments demonstrate that HSD yields consistent improvements in acceptance rates across diverse model families and benchmarks. Moreover, its strong explainability and generality make it readily integrable into a wide range of speculative decoding frameworks. Notably, integrating HSD into EAGLE-3 yields over a 12% performance gain, establishing state-of-the-art decoding efficiency without compromising distribution fidelity. Code is available at https://github.com/ZhouYuxuanYX/Hierarchical-Speculative-Decoding.

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.

Large language models have driven significant progress in natural language processing, but their deployment requires substantial compute and memory resources. As models scale, compression techniques become essential for balancing model quality with computational efficiency. Structured pruning, which removes less critical components of the model, is a promising strategy for reducing complexity. However, one-shot pruning often results in significant quality degradation, particularly in tasks requiring multi-step reasoning. To recover lost quality, supervised fine-tuning (SFT) is commonly applied, but it can lead to catastrophic forgetting by shifting the model's learned data distribution. Therefore, addressing the degradation from both pruning and SFT is essential to preserve the original model's quality. In this work, we utilize self-data distilled fine-tuning to address these challenges. Our approach leverages the original, unpruned model to generate a distilled dataset that preserves semantic richness and mitigates catastrophic forgetting by maintaining alignment with the base model's knowledge. Empirically, we demonstrate that self-data distillation consistently outperforms standard SFT, improving average accuracy by up to 8% on the HuggingFace OpenLLM Leaderboard v1. Specifically, when pruning six decoder blocks on Llama3.1-8B Instruct (i.e., 32 to 26 layers, reducing the model size from 8.03B to 6.72B parameters), our method retains 91.2% of the original model's accuracy compared to 81.7% with SFT, while reducing real-world FLOPs by 16.3%. Furthermore, combining self-data distilled models through model merging yields enhanced quality retention. Additionally, leveraging these pruned models in speculative decoding increases token acceptance rates, thereby improving inference efficiency in applied settings.

Language models have demonstrated remarkable capabilities in reasoning tasks through test-time scaling techniques like best-of-N sampling and tree search. However, these approaches often demand substantial computational resources, creating a critical trade-off between performance and efficiency. We introduce STAND (STochastic Adaptive N-gram Drafting), a novel model-free speculative decoding approach that leverages the inherent redundancy in reasoning trajectories to achieve significant acceleration without compromising accuracy. Our analysis reveals that reasoning paths frequently reuse similar reasoning patterns, enabling efficient model-free token prediction without requiring separate draft models. By introducing stochastic drafting and preserving probabilistic information through a memory-efficient logit-based N-gram module, combined with optimized Gumbel-Top-K sampling and data-driven tree construction, STAND significantly improves token acceptance rates. Extensive evaluations across multiple models and reasoning tasks (AIME-2024, GPQA-Diamond, and LiveCodeBench) demonstrate that STAND reduces inference latency by 60-65% compared to standard autoregressive decoding while maintaining accuracy. Furthermore, STAND outperforms state-of-the-art speculative decoding methods by 14-28% in throughput and shows strong performance even in single-trajectory scenarios, reducing inference latency by 48-58%. As a model-free approach, STAND can be applied to any existing language model without additional training, being a powerful plug-and-play solution for accelerating language model reasoning.

As Large Language Models (LLMs) scale up, inference efficiency becomes a critical bottleneck. Multi-Token Prediction (MTP) could accelerate LLM inference by predicting multiple future tokens in parallel. However, existing MTP approaches still face two challenges: limited acceptance rates of MTP heads, and difficulties in jointly training multiple MTP heads. Therefore, we propose MTP-D, a simple yet effective self-distillation method with minimal additional training cost, which boosts MTP head acceptance rates (+7.5\%) while maximumly preserving main-head performance. We also introduce a looped extension strategy for MTP-D, enabling effective and economical MTP head extension and further significant inference speedup to 1-head MTP (+220.4\%). Moreover, we systematically explore and validate key insights on the distillation strategies and the potential scalability of MTP through extensive experiments on seven benchmarks. These results demonstrate that our MTP-D and looped extension strategy effectively enhance MTP-head performance and inference efficiency, facilitating the practical usage of MTP in LLMs.

Speculative decoding, in which a lightweight draft model first generates a draft sequence that is then verified in parallel by the target model, has become a prevalent paradigm for accelerating large language model inference. Recent work such as DFlash further boosts drafting efficiency by leveraging diffusion drafters, whose parallel denoising mechanism enables draft generation in a single forward pass. In this work, we uncover a central pitfall of diffusion drafters: bidirectional attention is a double-edged sword. On one hand, it endows the model with parallel generation and global contextual modeling capabilities; on the other hand, this inherent global dependency introduces high variance at both the domain-level and the token-level: acceptance rates fluctuate substantially across different domains, and draft token quality also varies heterogeneously at different token positions. To tackle this issue, we propose AdaFlash framework, comprising two components: (i) an on-policy distillation (OPD) algorithm with reverse-KL divergence tailored for diffusion drafters, bringing stable convergence and effectively reducing domain-level variance; and (ii) an adaptive length head that dynamically adjusts the candidate sequence length on the fly, substantially lowering the verification cost of the target model and handling token-level variance. Experiments demonstrate that AdaFlash consistently improves speedup rate during deployment, with especially significant gains in high-concurrency scenarios, achieving up to approximately 66% higher throughput than previous state-of-the-art methods.

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.

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 accelerates autoregressive speech generation by letting a fast draft model propose tokens that a larger target model verifies. However, for speech LLMs that generate acoustic tokens, exact token matching is overly restrictive: many discrete tokens are acoustically or semantically interchangeable, reducing acceptance rates and limiting speedups. We introduce Principled Coarse-Graining (PCG), which verifies proposals at the level of Acoustic Similarity Groups (ASGs) derived from the target model's embedding space. By splitting each token's probability mass across the overlapping groups that contain it, we define an overlap-aware coarse-grained distribution and perform rejection sampling on the resulting group variable. This yields an exactness guarantee at the group level while allowing the accepted draft token to stand in for any member of the group in practice. On LibriTTS, PCG increases acceptance and throughput relative to standard speculative decoding and prior speech-specific relaxations while maintaining intelligibility and speaker similarity. These results suggest acoustically aware, group-level acceptance as a simple and general way to accelerate speech token generation while maintaining speech quality.

Vision-Language Models (VLMs) have revolutionized document parsing by enabling end-to-end mapping from images to structured text, imposing a significant latency bottleneck, particularly for token-dense documents. While Multi-Token Prediction (MTP) has emerged as a promising approach for accelerating inference, its potential is constrained by optimization instability when scaling to deeper look-ahead depth. In this paper, we propose P-MTP, a framework that leverages Progressive Multi-Token Prediction with a lightweight MTP module to scale the look-ahead depth for high-throughput document parsing. Specifically, we introduce Progressive Curriculum Loss that adaptively re-weights different look-ahead depths using cumulative path reliability and retrospective target consistency. By effectively suppressing gradient noise in long-range predictions, P-MTP, facilitates an automated easy-to-hard optimization transition, enabling the model to master increasingly distant look-ahead depths. Furthermore, we propose Confidence-Gated Dynamic Drafting to maximize the effective look-ahead depth and acceptance rate by adaptively calibrating speculative length during inference, thereby minimizing computational waste and further pushing the boundaries of inference speedup. Experimental results across multiple benchmarks and architectures demonstrate that P-MTP, achieves up to a 5times speedup with negligible loss in accuracy, providing the first successful validation of extensive look-ahead MTP in the document parsing domain.

As large language models (LLMs) become increasingly powerful, the sequential nature of autoregressive generation creates a fundamental throughput bottleneck that limits the practical deployment. While Multi-Token Prediction (MTP) has demonstrated remarkable benefits for model training efficiency and performance, its inherent potential for inference acceleration remains largely unexplored. This paper introduces FastMTP, a simple yet effective method that improves multi-step draft quality by aligning MTP training with its inference pattern, significantly enhancing speculative decoding performance. Our approach fine-tunes a single MTP head with position-shared weights on self-distilled data, enabling it to capture dependencies among consecutive future tokens and maintain high acceptance rates across multiple recursive draft steps. By integrating language-aware dynamic vocabulary compression into the MTP head, we further reduce computational overhead in the drafting process. Experimental results across seven diverse benchmarks demonstrate that FastMTP achieves an average of 2.03x speedup compared to standard next token prediction with lossless output quality, outperforming vanilla MTP by 82%. FastMTP requires only lightweight training and seamlessly integrates with existing inference frameworks, offering a practical and rapidly deployable solution for accelerating LLM inference.

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.

Reinforcement learning (RL) has become a key component in modern large language models, yet the rollout stage remains the key bottleneck in RL training pipelines. Although Multi-Token Prediction (MTP) offers a natural solution to accelerate rollouts through speculative decoding, many studies have observed that MTP acceptance rates degrade significantly during RL training, leading to limited speedup performance. To address this bottleneck, we present Bebop, a systematic study of MTP in LLM post-training, and offer practical recipes to integrate MTP into large-scale RL pipelines. First, we reveal that the MTP acceptance rate is fundamentally bounded by the fluctuation of model entropy, which demonstrates a clear negative linear relationship with the rise of entropy in the RL stage. Second, we show that probabilistic rejection sampling largely alleviates the disturbance introduced by entropy in RL compared to greedy draft sampling. We further identify that the conventional MTP training objectives (cross-entropy or KL) are suboptimal in such settings, and therefore we propose a novel end-to-end TV loss that directly optimizes multi-step rejection sampling acceptance rate, yielding ~10% acceptance rate improvements, achieving up to 95% acceptance rates and up to 25% extra inference throughput gains across mathematical reasoning, code generation, and agentic tasks. Third, we test various online MTP training strategies during RL and show that pre-RL MTP training with e2e TV loss and rejection sampling achieves a consistent acceptance rate and speedup throughout the entire RL, eliminating the need for costly online MTP updating. We provide extensive experiments and analysis that validate our findings. Experimental results show our method achieves up to 1.8x end-to-end acceleration in async RL training of Qwen3.5, Qwen3.6, and Qwen3.7 models.

Speculative decoding accelerates large language model inference by drafting multiple candidate tokens and verifying them in a single forward pass. Candidates are organized as a tree: deeper trees accept more tokens per step, but adding depth requires sacrificing breadth (fallback options) under a fixed verification budget. Existing training-free methods draft from a single token source and shape their trees without distinguishing candidate quality across origins. We observe that two common training-free token sources - n-gram matches copied from the input context, and statistical predictions from prior forward passes - differ dramatically in acceptance rate (~6x median gap, range 2-18x across five models and five benchmarks). We prove that when such a quality gap exists, the optimal tree is anisotropic (asymmetric): reliable tokens should form a deep chain while unreliable tokens spread as wide branches, breaking through the depth limit of balanced trees. We realize this structure in GOOSE, a training-free framework that builds an adaptive spine tree - a deep chain of high-acceptance context-matched tokens with wide branches of low-acceptance alternatives at each node. We prove that the number of tokens accepted per step is at least as large as that of either source used alone. On five LLMs (7B-33B) and five benchmarks, GOOSE achieves 1.9-4.3x lossless speedup, outperforming balanced-tree baselines by 12-33% under the same budget.

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%.

The auto-regressive decoding of Large Language Models (LLMs) results in significant overheads in their hardware performance. While recent research has investigated various speculative decoding techniques for multi-token generation, these efforts have primarily focused on improving processing speed such as throughput. Crucially, they often neglect other metrics essential for real-life deployments, such as memory consumption and training cost. To overcome these limitations, we propose a novel parallel prompt decoding that requires only 0.0002% trainable parameters, enabling efficient training on a single A100-40GB GPU in just 16 hours. Inspired by the human natural language generation process, PPD approximates outputs generated at future timesteps in parallel by using multiple prompt tokens. This approach partially recovers the missing conditional dependency information necessary for multi-token generation, resulting in up to a 28% higher acceptance rate for long-range predictions. Furthermore, we present a hardware-aware dynamic sparse tree technique that adaptively optimizes this decoding scheme to fully leverage the computational capacities on different GPUs. Through extensive experiments across LLMs ranging from MobileLlama to Vicuna-13B on a wide range of benchmarks, our approach demonstrates up to 2.49times speedup and maintains a minimal runtime memory overhead of just 0.0004%. More importantly, our parallel prompt decoding can serve as an orthogonal optimization for synergistic integration with existing speculative decoding, showing up to 1.22times further speed improvement. Our code is available at https://github.com/hmarkc/parallel-prompt-decoding.

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.

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.

The escalating demand for efficient decoding in large language models (LLMs) is particularly critical for reasoning-intensive architectures like OpenAI-o3 and DeepSeek-R1, which depend on extended chain-of-thought reasoning. This study investigates speculative decoding techniques through dense LLM architectures to establish foundational insights for accelerating reasoning tasks. While speculative decoding methods leveraging parallel draft-verification cycles have emerged as promising acceleration techniques, the scaling laws governing decoding efficiency remain under-explored compared to conventional backbone LLMs developed through Pretraining->SFT->RLHF training paradigms. In this work, we discover Log-linear Scaling Laws (Theorem 1.1, 1.2 and 1.3) governing draft model acceptance rate (or decoding speed) across three dimensions: pretraining token volume, draft model capacity, and decoding batch size. Building on these laws, we achieve Scylla, which coordinates multi-dimensional scaling for popular LLMs (Llama2/3, Qwen2.5). Empirical validation shows Scylla achieves 1.5-2.2 higher acceptance rate than EAGLE2 and 0.3 higher than EAGLE3 at temperature T = 0, with peak performance gains on summarization and QA tasks (Figure 2). Industrial inference engine deployments demonstrate 2X decoding throughput improvements over EAGLE2 (Table 5), validating the transformative potential of systematic scaling for efficient LLM inference. Code will be released later.

We introduce Nemotron-Labs-Diffusion, a tri-mode language model (LM) that unifies AR, diffusion, and self-speculation decoding within a single architecture. Trained with a joint AR-diffusion objective, Nemotron-Labs-Diffusion can switch modes to sustain high throughput across deployment settings and concurrency levels. Our study shows that (1) AR and diffusion objectives are complementary: diffusion improves lookahead planning, while AR provides left-to-right linguistic priors. (2) In self-speculation mode, diffusion drafts while AR verifies, outperforming multi-token prediction (MTP) methods in both acceptance rate and real-device efficiency. (3) A speed-of-light analysis further demonstrates diffusion's long-term potential, with up to 76.5% more tokens per forward pass than self-speculation under an optimal sampler. Scaling to 3B, 8B, and 14B parameters, our Nemotron-Labs-Diffusion family, including base, instruct, and vision-language models, consistently outperforms state-of-the-art open-source AR and diffusion LMs in both accuracy and speed. For example, Nemotron-Labs-Diffusion-8B decodes 6x more tokens per forward than Qwen3-8B with comparable accuracy, translating to 4x higher throughput on SPEED-Bench with SGLang on a GB200 GPU.

Model quantization has become essential for efficient large language model deployment, yet existing approaches present clear trade-offs: methods such as GPTQ and AWQ achieve practical compression but are lossy, while lossless techniques preserve fidelity but lack inference acceleration. This paper explores the middle ground of statistically-lossless compression, examining three complementary aspects of what losslessness means for quantized LLMs. First, task-lossless compression preserves zero-shot benchmark accuracy within natural sampling variance and is achievable at aggressive bitwidths. Second, we formalize the stricter notion of distribution-lossless compression, requiring the quantized model's next-token distribution to be practically indistinguishable from the original, and propose the Expected Acceptance Rate (EAR), the maximum token-agreement probability under optimal coupling, as a directly interpretable fidelity metric. For example, EAR >= 0.99 means 99% agreement. Third, we prove a gamma-squared variance law showing that symmetric quantization inflates noise variance by gamma^2 relative to asymmetric quantization, making asymmetric quantization a prerequisite for distribution-lossless fidelity but not for task-level preservation. Through SLQ, a layer-wise non-uniform method with asymmetric quantization and wide bitwidth search, we obtain task-lossless compression at well below 4 bits per parameter, as low as 3.3 bits depending on the model, distribution-lossless compression at 5-6 bits per parameter on average, and inference speedups of 1.7-3.7x compared to FP16 using optimized kernels. Source code is available at [https://github.com/IST-DASLab/SLQ](https://github.com/IST-DASLab/SLQ).

Inference of Large Language Models (LLMs) across computer clusters has become a focal point of research in recent times, with many acceleration techniques taking inspiration from CPU speculative execution. These techniques reduce bottlenecks associated with memory bandwidth, but also increase end-to-end latency per inference run, requiring high speculation acceptance rates to improve performance. Combined with a variable rate of acceptance across tasks, speculative inference techniques can result in reduced performance. Additionally, pipeline-parallel designs require many user requests to maintain maximum utilization. As a remedy, we propose PipeInfer, a pipelined speculative acceleration technique to reduce inter-token latency and improve system utilization for single-request scenarios while also improving tolerance to low speculation acceptance rates and low-bandwidth interconnects. PipeInfer exhibits up to a 2.15times improvement in generation speed over standard speculative inference. PipeInfer achieves its improvement through Continuous Asynchronous Speculation and Early Inference Cancellation, the former improving latency and generation speed by running single-token inference simultaneously with several speculative runs, while the latter improves speed and latency by skipping the computation of invalidated runs, even in the middle of inference.

Speculative Decoding (SD) accelerates low-concurrency LLM inference by employing a draft-then-verify paradigm. However, mainstream methods typically rely on multi-token prediction, which introduces escalating prediction difficulty and serial drafting latency. To address these, we propose Speculative Pipeline Decoding (SPD), a groundbreaking framework that unlocks the true potential of pipeline parallelism. By partitioning the target LLM into n pipeline stages, SPD allows LLM to process n tokens in parallel to accelerate decoding. To continuous fill the pipeline in single sequence decoding, a speculation module aggregates intermediate features across different pipeline depths to predict the next token, executing strictly in parallel with the target model's pipeline step, to realize bounded difficulty, higher acceptance rates, and zero latency bubbles. Our experiments demonstrate that SPD achieves a significantly higher theoretical speedup compared to mainstream baselines, offering a highly scalable solution for LLM decoding acceleration. Our code is available at https://github.com/yuyijiong/speculative_pipeline_decoding

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.

Large language models have shown impressive capabilities across a variety of NLP tasks, yet their generating text autoregressively is time-consuming. One way to speed them up is speculative decoding, which generates candidate segments (a sequence of tokens) from a fast draft model that is then verified in parallel by the target model. However, the acceptance rate of candidate tokens receives limitations from several factors, such as the model, the dataset, and the decoding setup. This paper proposes sampling multiple candidates from a draft model and then organising them in batches for verification. We design algorithms for efficient multi-candidate verification while maintaining the distribution of the target model. Our approach shows significant improvements in acceptance rates on multiple datasets and models, consistently outperforming standard speculative decoding.

Speculative decoding (SD) accelerates Large Language Model (LLM) generation by using an efficient draft model to propose the next few tokens, which are verified by the LLM in a single forward call, reducing latency while preserving its outputs. We focus on retrieval-based SD where the draft model retrieves the next tokens from a non-parametric datastore. Sparse retrieval (REST), which operates on the surface form of strings, is currently the dominant paradigm due to its simplicity and scalability. However, its effectiveness is limited due to the usage of short contexts and exact string matching. Instead, we introduce Dense Retrieval for Speculative Decoding (DReSD), a novel framework that uses approximate nearest neighbour search with contextualised token embeddings to retrieve the most semantically relevant token sequences for SD. Extensive experiments show that DReSD achieves (on average) 87% higher acceptance rates, 65% longer accepted tokens and 19% faster generation speeds compared to sparse retrieval (REST).

Inference with modern Large Language Models (LLMs) is expensive and time-consuming, and speculative sampling has proven to be an effective solution. Most speculative sampling methods such as EAGLE use a static draft tree, implicitly assuming that the acceptance rate of draft tokens depends only on their position. Interestingly, we found that the acceptance rate of draft tokens is also context-dependent. In this paper, building upon EAGLE, we propose EAGLE-2, which introduces a new technique of context-aware dynamic draft tree into drafting modeling. This improvement leverages the fact that the draft model of EAGLE is well-calibrated: the confidence scores from the draft model approximate acceptance rates with small errors. We conducted extensive evaluations on three series of LLMs and six tasks, with EAGLE-2 achieving speedup ratios 3.05x-4.26x, which is 20%-40% faster than EAGLE-1. EAGLE-2 also ensures that the distribution of the generated text remains unchanged, making it a lossless acceleration algorithm.

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.

Autoregressive large language models (LLMs) deliver strong performance but require inherently sequential decoding, leading to high inference latency and poor GPU utilization. Speculative decoding mitigates this bottleneck by using a fast draft model whose outputs are verified in parallel by the target LLM; however, existing methods still rely on autoregressive drafting, which remains sequential and limits practical speedups. Diffusion LLMs offer a promising alternative by enabling parallel generation, but current diffusion models typically underperform compared with autoregressive models. In this paper, we introduce DFlash, a speculative decoding framework that employs a lightweight block diffusion model for parallel drafting. By generating draft tokens in a single forward pass and conditioning the draft model on context features extracted from the target model, DFlash enables efficient drafting with high-quality outputs and higher acceptance rates. Experiments show that DFlash achieves over 6x lossless acceleration across a range of models and tasks, delivering up to 2.5x higher speedup than the state-of-the-art speculative decoding method EAGLE-3.

Recent advancements in speculative decoding have demonstrated considerable speedup across a wide array of large language model (LLM) tasks. Speculative decoding inherently relies on sacrificing extra memory allocations to generate several candidate tokens, of which acceptance rate drives the speedup. However, deploying speculative decoding on memory-constrained devices, such as mobile GPUs, remains as a significant challenge in real-world scenarios. In this work, we present a device-aware inference engine named SpecMemo that can smartly control memory allocations at finer levels to enable multi-turn chatbots with speculative decoding on such limited memory devices. Our methodology stems from theoretically modeling memory footprint of speculative decoding to determine a lower bound on the required memory budget while retaining speedup. SpecMemo empirically acquires a careful balance between minimizing redundant memory allocations for rejected candidate tokens and maintaining competitive performance gains from speculation. Notably, with SpecMemo's memory management, we maintain 96% of overall throughput from speculative decoding on MT-Bench, with reduced generation-memory by 65% on single Nvidia Titan RTX. Given multiple constrained GPUs, we build on top of previous speculative decoding architectures to facilitate big-model inference by distributing Llama-2-70B-Chat model, on which we provide novel batched speculative decoding to increase usability of multiple small server GPUs. This novel framework demonstrates 2x speedup over distributed and batched vanilla decoding with the base model on eight AMD MI250 GPUs. Moreover, inference throughput increases remarkably 8x with batch size 10. Our work contributes to democratized LLM applications in resource-constrained environments, providing a pathway for faster and cheaper deployment of real-world LLM applications with robust performance.

Speculative decoding's effectiveness is governed by two quantities: the acceptance rate of drafted tokens and the cost of producing them. We show the two objectives can be co-optimized with algorithm-system co-design, and present JetSpec++, which retains a strong learned draft model while hiding its cost almost entirely. JetSpec++ combines three components: intra-block self refinement, which trains the draft model to iteratively correct its own parallel predictions within a block of tokens; inter-block foresight drafting, which begins the next block before the current one is verified; and pipelined drafting and verification, which overlaps the two stages so drafting's marginal wall-clock cost approaches zero. In comparison with state-of-the-art speculative decoding methods, JetSpec++ attains a wall-clock speedup of up to 7.45times over autoregressive (AR) decoding and 1.54times over DSpark, with up to 1,361 tokens per second for a single request on GB200, and an average acceptance length of 8.94 across seven benchmarks, while preserving the target model's output distribution. Our code and models are available at https://github.com/hao-ai-lab/JetSpec.

Leveraging recent advancements in large language models, modern neural code completion models have demonstrated the capability to generate highly accurate code suggestions. However, their massive size poses challenges in terms of computational costs and environmental impact, hindering their widespread adoption in practical scenarios. Dynamic inference emerges as a promising solution, as it allocates minimal computation during inference while maintaining the model's performance. In this research, we explore dynamic inference within the context of code completion. Initially, we conducted an empirical investigation on GPT-2, focusing on the inference capabilities of intermediate layers for code completion. We found that 54.4% of tokens can be accurately generated using just the first layer, signifying significant computational savings potential. Moreover, despite using all layers, the model still fails to predict 14.5% of tokens correctly, and the subsequent completions continued from them are rarely considered helpful, with only a 4.2% Acceptance Rate. These findings motivate our exploration of dynamic inference in code completion and inspire us to enhance it with a decision-making mechanism that stops the generation of incorrect code. We thus propose a novel dynamic inference method specifically tailored for code completion models. This method aims not only to produce correct predictions with largely reduced computation but also to prevent incorrect predictions proactively. Our extensive evaluation shows that it can averagely skip 1.7 layers out of 16 layers in the models, leading to an 11.2% speedup with only a marginal 1.1% reduction in ROUGE-L.

Striking an optimal balance between minimal drafting latency and high speculation accuracy to enhance the inference speed of Large Language Models remains a significant challenge in speculative decoding. In this paper, we introduce Falcon, an innovative semi-autoregressive speculative decoding framework fashioned to augment both the drafter's parallelism and output quality. Falcon incorporates the Coupled Sequential Glancing Distillation technique, which fortifies inter-token dependencies within the same block, leading to increased speculation accuracy. We offer a comprehensive theoretical analysis to illuminate the underlying mechanisms. Additionally, we introduce a Custom-Designed Decoding Tree, which permits the drafter to generate multiple tokens in a single forward pass and accommodates multiple forward passes as needed, thereby boosting the number of drafted tokens and significantly improving the overall acceptance rate. Comprehensive evaluations on benchmark datasets such as MT-Bench, HumanEval, and GSM8K demonstrate Falcon's superior acceleration capabilities. The framework achieves a lossless speedup ratio ranging from 2.91x to 3.51x when tested on the Vicuna and LLaMA2-Chat model series. These results outstrip existing speculative decoding methods for LLMs, including Eagle, Medusa, Lookahead, SPS, and PLD, while maintaining a compact drafter architecture equivalent to merely two Transformer layers.

Agentic AI systems built on large language models (LLMs) offer significant potential for automating complex workflows, from software development to customer support. However, LLM agents often underperform due to suboptimal configurations; poorly tuned prompts, tool descriptions, and parameters that typically require weeks of manual refinement. Existing optimization methods either are too complex for general use or treat components in isolation, missing critical interdependencies. We present ARTEMIS, a no-code evolutionary optimization platform that jointly optimizes agent configurations through semantically-aware genetic operators. Given only a benchmark script and natural language goals, ARTEMIS automatically discovers configurable components, extracts performance signals from execution logs, and evolves configurations without requiring architectural modifications. We evaluate ARTEMIS on four representative agent systems: the ALE Agent for competitive programming on AtCoder Heuristic Contest, achieving a 13.6% improvement in acceptance rate; the Mini-SWE Agent for code optimization on SWE-Perf, with a statistically significant 10.1\% performance gain; and the CrewAI Agent for cost and mathematical reasoning on Math Odyssey, achieving a statistically significant 36.9% reduction in the number of tokens required for evaluation. We also evaluate the MathTales-Teacher Agent powered by a smaller open-source model (Qwen2.5-7B) on GSM8K primary-level mathematics problems, achieving a 22\% accuracy improvement and demonstrating that ARTEMIS can optimize agents based on both commercial and local models.

Large language models (LLMs) remain susceptible to jailbreak and direct prompt-injection attacks, yet the strongest defensive filters frequently over-refuse benign queries and degrade user experience. Previous work on jailbreak and prompt injection detection such as GradSafe, detects unsafe prompts with a single "accept all" anchor token, but its threshold is brittle and it offers no deterministic guarantee that harmful content will not be emitted once decoding begins. We introduce Gradient-Controlled Decoding (GCD), a training-free guardrail that combines an acceptance anchor token ("Sure") and refusal anchor token ("Sorry") tightening the decision boundary and significantly lowering false positives. In the mitigation stage, if a prompt is flagged, GCD preset-injects one or two refusal tokens ("Sorry, I can't...") before autoregressive decoding resumes, guaranteeing first-token safety regardless of sampling strategy. On ToxicChat, XSTest-v2, and AdvBench, GCD reduces false positives by 52% vs. GradSafe at comparable recall, lowers attack success rate by up to 10% vs. the strongest decoding-only baseline, adds under 15-20 ms latency on an average on V100 instances, transfers to LLaMA-2-7B, Mixtral-8x7B, and Qwen-2-7B, and requires only 20 demonstration templates.

Frontier deep research agents (DRAs) plan a research task, synthesize across documents, and return a structured deliverable on demand. They are being deployed in enterprise workflows faster than they are being evaluated. Existing benchmarks measure factual recall, single-hop QA, or generic agentic skill, missing the multi-document, decision-grade work DRAs are deployed to produce. We introduce a benchmark targeting the structured analytical deliverables that fill a management consultant's typical week. We grade three frontier agents, namely Claude Opus 4.6 with web search, OpenAI o3-deep-research, and Google Gemini 3.1 Pro deep-research, on 42 SME-authored prompts. Each of the 126 responses is scored on two layers: deterministic ground-truth verifiers (mean 13.8 per task) and a five-criterion 0-3 SME rubric, composed into a Verifier-Rubric Score (VRS) on 0-100. Most prompts embed cognitive traps that penalize surface-pattern matching. Acceptance under our joint threshold (rubric mean >= 2.5 and verifier rate >= 80%) is uniformly low: Gemini 21.4%, o3 9.5%, Claude 9.5%. Mean VRS scores agree with published rubric-based benchmarks (our top 62.6 vs. APEX-v1 64.2, ProfBench 65.9, ResearchRubrics < 68%), validating the rubric construct. ACCEPT rates sit below APEX-Agents' MC-segment Pass@1 band (12.3-22.7%) on dedicated DR agents; our floor is three points lower despite the harness advantage, opened by stricter conjunctive grading and trap design. Each agent fails distinctively. Claude produces the deliverable most reliably (4.5x the others' rate on file-required tasks) but carries the highest fabrication signature. o3 has the cleanest reasoning average yet drops required sections and propagates arithmetic errors. Gemini is bimodal, with the highest acceptance rate alongside the most zero-scored rubric cells.

Autoregressive large language model inference repeatedly invokes the target model to generate one token at a time, making generation sensitive to GPU memory movement and sequential execution. This study evaluates two-token multi-token prediction (MTP) against autoregressive decoding in a controlled single-request deployment on an NVIDIA A10G GPU. A 360-request benchmark covered plain-text, reasoning-intensive, and tool-calling workloads, while runtime telemetry, Nsight Systems, PyTorch Profiler, and selected Nsight Compute measurements were used to explain the observed performance. MTP increased output throughput by \(1.91\times\) to \(2.19\times\) across all prompts and reduced time to first output by 10.0--14.2\%. Median mean acceptance length ranged from 2.370 to 2.595 tokens per verification iteration. Profiling showed that MTP introduced a longer and more complex execution path, including proposal, sampling, attention, gathering, and reduction operations. However, it required 56.4--78.1\% fewer executions of the selected repeating CUDA Graph per generated token. The dominant MTP GEMM kernel was not faster than the dominant autoregressive GEMV kernel, and selected instances of both approached the A10G memory-bandwidth limit. These results show that MTP improved inference through amortization: greater token progress reduced repeated GPU execution sufficiently to outweigh the additional speculative-execution cost.

The throughput gains of speculative decoding depend on the trade-off between acceptance length and execution latency. We ask when additional acceptance is worth the cost of obtaining it. We establish a comparison protocol spanning training, inference, and evaluation, remeasuring released checkpoints and retraining six drafter recipes on shared target-generated data across ten benchmarks and two target models. Alignment changes the picture: gaps that appear large in published comparisons shrink substantially under shared training and inference conditions. This narrowing motivates two deeper questions: how is acceptance improved at different costs, and when do these gains improve throughput? We find that acceptance can be improved through nearly free objective and context changes or more expensive token-to-token conditioning. DCore illustrates the former route as a strong baseline with little additional measured drafting overhead. The payoff is conditional: task uncertainty and serving conditions affect acceptance returns, execution costs, and the preferred drafter and budget. These findings inform future drafter design and deployment choices.

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.

Scaling laws enable the optimal selection of data amount and language model size, yet the impact of the data unit, the token, on this relationship remains underexplored. In this work, we systematically investigate how the information granularity of tokens, controlled by the compression rate (i.e., average bytes of text per token), affects scaling trends. We train 988 latent tokenized models (BLT) ranging from 50M to 7B parameters that enable setting the desired compression rate. This flexibility allows us to study the role of compression rate well beyond 4.57 bytes per token obtained with a popular BPE tokenizer. Our experiments reveal that in compute-optimal configurations, model parameter counts scale proportionally to data size measured in bytes, not in tokens as commonly perceived (Kaplan et al., 2020; Hoffmann et al., 2022). Furthermore, we discover that the optimal compression rate differs from the one obtained with BPE and decreases with compute. These findings generalize to both latent and subword tokenization, as well as to languages other than English, guiding language model developers on tokenization scheme selection for maximal compute efficiency.

Practicing conversations with large language models (LLMs) presents a promising alternative to traditional in-person language learning. However, most LLMs generate text at a near-native level of complexity, making them ill-suited for beginner learners (CEFR: A1-A2). In this paper, we investigate whether controllable generation techniques -- specifically modular methods that do not require model fine-tuning -- can adapt LLM outputs to better support absolute beginners. We evaluate these methods through both automatic metrics and a user study with university-level learners of Japanese. Our findings show that while prompting alone fails to control output difficulty, the use of future discriminators (Yang and Klein, 2021) significantly improves output comprehensibility (from 40.4\% to 84.3\%). We further introduce a novel token-level evaluation metric, Token Miss Rate (TMR), that quantifies the proportion of incomprehensible tokens per utterance and correlates strongly with human judgments. To support future research in AI-assisted language learning, we release our code, models, annotation tools, and dataset.

Most language models deployed outside the United States and China are fine-tuned descendants of a small number of pretrained families. We argue that fine-tuning leaves the decisive properties of a model inherited from its base: vocabulary, knowledge, cultural alignment, memorized data and license. We contribute (i) an updated, reproducible measurement of the tokenization premium across 31 languages and 7 tokenizers released in 2023-2024, on 27 articles of the Universal Declaration of Human Rights, with explicit correction for unknown-token collapse. Premiums reach 10x for Amharic in older tokenizers and remain at 6.8x in the newest commercial one; Indigenous South American languages pay between 1.6x and 2.7x under global tokenizers. A vocabulary built from Portuguese lowers the premium for 11 of 13 Latin-script Global South languages relative to every global tokenizer tested, by about 24% for Guarani and Quechua. (ii) An eight-test black-box sovereignty audit that a buyer can run without access to weights. (iii) A regime argument: under energy and liability constraints, accuracy per joule, calibration and selective risk should replace raw accuracy as design objectives, with metrics that abstention cannot inflate. We close with preliminary internal measurements from a proprietary architecture built around these objectives, reported with their limitations. Data and measurement code are included.

Frontier LLMs are increasingly utilised across academia, society and industry. A commonly used unit for comparing models, their inputs and outputs, and estimating inference pricing is the token. In general, tokens are used as a stable currency, assumed to be broadly consistent across tokenizers and contexts, enabling direct comparisons. However, tokenization varies significantly across models and domains of text, making naive interpretation of token counts problematic. We quantify this variation by providing a comprehensive empirical analysis of tokenization, exploring the compression of sequences to tokens across different distributions of textual data. Our analysis challenges commonly held heuristics about token lengths, finding them to be overly simplistic. We hope the insights of our study add clarity and intuition toward tokenization in contemporary LLMs.

The wide adoption of AI agents in complex human workflows is driving rapid growth in LLM token consumption. When agents are deployed on tasks that require a significant amount of tokens, three questions naturally arise: (1) Where do AI agents spend the tokens? (2) Which models are more token-efficient? and (3) Can agents predict their token usage before task execution? In this paper, we present the first systematic study of token consumption patterns in agentic coding tasks. We analyze trajectories from eight frontier LLMs on SWE-bench Verified and evaluate models' ability to predict their own token costs before task execution. We find that: (1) agentic tasks are uniquely expensive, consuming 1000x more tokens than code reasoning and code chat, with input tokens rather than output tokens driving the overall cost; (2) token usage is highly variable and inherently stochastic: runs on the same task can differ by up to 30x in total tokens, and higher token usage does not translate into higher accuracy; instead, accuracy often peaks at intermediate cost and saturates at higher costs; (3) models vary substantially in token efficiency: on the same tasks, Kimi-K2 and Claude-Sonnet-4.5, on average, consume over 1.5 million more tokens than GPT-5; (4) task difficulty rated by human experts only weakly aligns with actual token costs, revealing a fundamental gap between human-perceived complexity and the computational effort agents actually expend; and (5) frontier models fail to accurately predict their own token usage (with weak-to-moderate correlations, up to 0.39) and systematically underestimate real token costs. Our study offers new insights into the economics of AI agents and can inspire future research in this direction.

Speculative Decoding (SD) is a technique to accelerate the inference of Large Language Models (LLMs) by using a lower complexity draft model to propose candidate tokens verified by a larger target model. To further improve efficiency, Multi-Candidate Speculative Decoding (MCSD) improves upon this by sampling multiple candidate tokens from the draft model at each step and verifying them in parallel, thus increasing the chances of accepting a token and reducing generation time. Existing MCSD methods rely on the draft model to initialize the multi-candidate sequences and use static length and tree attention structure for draft generation. However, such an approach suffers from the draft and target model's output distribution differences, especially in dynamic generation context. In this work, we introduce an improved version of MCSD that includes a target model initialized multi-candidate process, dynamic sliced topology-aware causal mask for dynamic length adjustment, and decision models to optimize early stopping. Our framework improves the acceptance rate, defined as the ratio of the longest draft sequence length accepted by the target model over the maximum draft sequence length, by a maximum of 164% and gains a maximum of 75% generation speed up over the MCSD baseline. We also conduct an ablation study to evaluate the impact of the decision model.

This paper measures hour-of-day effects, filter precision, fragility, and realised yield in a 15-day paper-traded deployment of an autonomous memecoin trading system on Solana decentralised exchanges. The 190-trade sample (March 29 to April 12, 2026) shows a 40.5 percent win rate, mean per-trade return of +0.62 percent, cumulative +117.7 percent (net SOL +0.039), skewness -1.21, excess kurtosis 6.61. A Mann-Whitney U test of three poorest-performing UTC hours (2, 13, 23) against the others yields U = 1,274, p = 0.22; directional but not significant at n = 190. The three hours were selected in-sample, so the comparison is exploratory, not confirmatory. A parallel counterfactual rejection-tracking system collected 4,874 forward-sample observations across 184 distinct rejection events. Of those events, 17.9 percent reached a 50 percent drawdown from reference within 24 hours; 26.0 percent of forward samples recorded the rejected token below half-reference. The filter stack avoided these realised drawdowns: evidence that the rejection criteria are net-positive against forward-market outcomes. Fragility is the principal caveat. Removing the top three trades (1.6 percent of sample) flips cumulative return unprofitable. Profitability rests on a small number of large winners and is structurally fragile. The dataset and audit script are deposited under CC-BY-4.0 (Zenodo DOI 10.5281/zenodo.20043302).

This paper presents a dynamic cryptocurrency portfolio optimization strategy that integrates technical indicators and sentiment analysis to enhance investment decision-making. The proposed method employs the 14-day Relative Strength Index (RSI) and 14-day Simple Moving Average (SMA) to capture market momentum, while sentiment scores are extracted from news articles using the VADER (Valence Aware Dictionary and sEntiment Reasoner) model, with compound scores quantifying overall market tone. The large language model Google Gemini is used to further verify the sentiment scores predicted by VADER and give investment decisions. These technical indicator and sentiment signals are incorporated into the expected return estimates before applying mean-variance optimization with constraints on asset weights. The strategy is evaluated through a rolling-window backtest over cryptocurrency market data, with Bitcoin (BTC) and an equal-weighted portfolio of selected cryptocurrencies serving as benchmarks. Experimental results show that the proposed approach achieves a cumulative return of 38.72, substantially exceeding Bitcoin's 8.85 and the equal-weighted portfolio's 21.65 over the same period, and delivers a higher Sharpe ratio (1.1093 vs. 0.8853 and 1.0194, respectively). However, the strategy exhibits a larger maximum drawdown (-18.52%) compared to Bitcoin (-4.48%) and the equal-weighted portfolio (-11.02%), indicating higher short-term downside risk. These results highlight the potential of combining sentiment and technical signals to improve cryptocurrency portfolio performance, while also emphasizing the need to address risk exposure in volatile markets.

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.

Language models have graduated from being research prototypes to commercialized products offered as web APIs, and recent works have highlighted the multilingual capabilities of these products. The API vendors charge their users based on usage, more specifically on the number of ``tokens'' processed or generated by the underlying language models. What constitutes a token, however, is training data and model dependent with a large variance in the number of tokens required to convey the same information in different languages. In this work, we analyze the effect of this non-uniformity on the fairness of an API's pricing policy across languages. We conduct a systematic analysis of the cost and utility of OpenAI's language model API on multilingual benchmarks in 22 typologically diverse languages. We show evidence that speakers of a large number of the supported languages are overcharged while obtaining poorer results. These speakers tend to also come from regions where the APIs are less affordable to begin with. Through these analyses, we aim to increase transparency around language model APIs' pricing policies and encourage the vendors to make them more equitable.

Commercial large language models bill, scale latency, and budget context per token. Yet tokenizers assign more subword tokens to the same meaning in some languages than in others, so speakers of languages with high token-fertility pay a structural penalty before a model is ever invoked. This penalty is documented for multilingual settings in general, but it has not been measured systematically for African languages at the level of enterprise deployment economics and cognitive context capacity. We measure it across 20 African languages spanning five language families and three scripts (Latin, Ge'ez/Ethiopic, N'Ko; 19 appear in the primary FLORES-200+ corpus, with Nigerian Pidgin measured via MAFAND-MT only), using parallel corpora so that the language effect is isolated from content. Across 11 frontier and open tokenizers on FLORES-200+, every African language carries a tokenization premium above English (median 1.88x on GPT-5 / o200k_base, up to 8.92x for N'Ko); the penalty is largest for Ethiopic and N'Ko scripts (reaching 7-9x) and is near-invariant across corpora (FLORES vs SIB-200 Pearson r = 0.9998). Translated into deployment terms, this results in up to 8.9x inference cost and an equivalent generation-latency multiplier (N'Ko vs English on GPT-5; 7.4x for Amharic), and as little as 11% of English's effective context window. The best currently available tokenizer for African languages, Gemma 4, reduces the mean premium from 3.31x (cl100k_base) to 2.38x, but no tokenizer eliminates the penalty. We release an open measurement tool (afri-fertility), a public leaderboard, a results dataset, and mitigation guidance for African builders. The penalty falls hardest on the languages whose speakers can least afford it, a digital divide encoded directly into the subword vocabulary.

We study reliability in autonomous language-model agents that translate user mandates into validated tool actions under real capital. The setting is DX Terminal Pro, a 21-day deployment in which 3,505 user-funded agents traded real ETH in a bounded onchain market. Users configured vaults through structured controls and natural-language strategies, but only agents could choose normal buy/sell trades. The system produced 7.5M agent invocations, roughly 300K onchain actions, about $20M in volume, more than 5,000 ETH deployed, roughly 70B inference tokens, and 99.9% settlement success for policy-valid submitted transactions. Long-running agents accumulated thousands of sequential decisions, including 6,000+ prompt-state-action cycles for continuously active agents, yielding a large-scale trace from user mandate to rendered prompt, reasoning, validation, portfolio state, and settlement. Reliability did not come from the base model alone; it emerged from the operating layer around the model: prompt compilation, typed controls, policy validation, execution guards, memory design, and trace-level observability. Pre-launch testing exposed failures that text-only benchmarks rarely measure, including fabricated trading rules, fee paralysis, numeric anchoring, cadence trading, and misread tokenomics. Targeted harness changes reduced fabricated sell rules from 57% to 3%, reduced fee-led observations from 32.5% to below 10%, and increased capital deployment from 42.9% to 78.0% in an affected test population. We show that capital-managing agents should be evaluated across the full path from user mandate to prompt, validated action, and settlement.

Large language models are increasingly deployed as general-purpose educational and technical assistance systems, but their underlying infrastructure does not treat languages equally. One underexamined source of disparity is tokenization: semantically equivalent content can require substantially different token counts across languages, affecting API cost, latency, and usable context length before a model is invoked. We introduce the Tokenization Equity Audit (TEA), a reproducible benchmark for measuring tokenization premiums in technical tutoring content. TEA evaluates three widely used tokenizers, GPT-4o's o200k base, Qwen2.5-7B, and Mistral-7B, on a 120-item Python debugging corpus translated from English into Bengali, Hindi, Arabic, Tamil, and Yoruba. Bengali and Hindi serve as the primary validated cases, while the remaining languages provide exploratory cross-script and cross-family comparisons. Across this corpus, Bengali requires (1.56\times) as many GPT-4o tokens as English, reducing a nominal 128k-token context window to an effective 82k-token English-equivalent capacity for the same semantic content. With the Qwen2.5 and Mistral tokenizers, Bengali requires up to (4.5\times) the English token count. Yoruba, despite using the Latin script, exhibits the highest GPT-4o tokenization premium at (2.37\times), indicating that tokenization inequity cannot be explained by script family alone. These results demonstrate that tokenization can create measurable economic and functional barriers, highlighting the need to treat tokenization as an equity-relevant infrastructure layer for underserved language communities, particularly where educational systems depend on low-cost or offline-capable AI tools.

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.

Software delegates more of its branches to models every year: which queue a ticket enters, whether a command is safe to run, whether a claim clears without a person. What the program needs back is not prose. It is one of n declared options and a number it can threshold. Today that costs a round trip to a frontier model -- hundreds of milliseconds, a per-token bill, and a parser -- for a question that is usually a conjunction of three clauses. this-that-model-1.0 is a 2B-parameter typed decision model. Its answer is read directly from the hidden state at a designated position and restricted to the option set the caller declared, so no text is generated, nothing can be malformed, and every question in a request is answered in the same forward pass. It decides in 30.9 ms on one laptop GPU and generates zero output tokens doing it, where a frontier API call costs 8758 ms and the hosted systems that answer these questions well spend between 21 and 212 generated tokens per question thinking first, billed for every one. It sustains 32 decisions per second on one consumer GPU and never lets the state leave the machine. On a third party's recorded cohort of 68 decision questions, on their inputs and their wording, it scores 0.941 with a Brier score of 0.042, against 0.765 and 0.133 for the hosted service Jev on the same items. One pass of our 42-family internal suite takes 32 seconds and 0.000217 USD of electricity; the most accurate hosted model we measured needs 155.2 minutes and 10.636 USD. We also report where it loses. On multi-step arithmetic, which a single forward pass cannot carry intermediate results through, it scores 0.560 against their 0.98 to 1.00, and a targeted second training round improved the five task families it was written for and transferred to none of the other 13. The model is open-sourced in https://huggingface.co/flock-io/this-that-model-1.0

Tokenization is a foundational step in NLP pipelines, yet standard evaluation metrics such as fertility rate fail to capture morphological correctness for agglutinative languages. We present QuechuaTok, a systematic benchmark comparing four tokenization strategies - BPE, Unigram LM, WordPiece, and a morphology-aware PRPE tokenizer - for Southern Quechua (quz), a low-resource agglutinative language spoken by 8-10 million people in South America. Using a 200k-sentence corpus and the SQUOIA finite-state morphological analyzer (Rios, 2016) as silver standard, we evaluate three metrics: fertility rate, OOV rate, and morphological boundary accuracy (MorphAcc). Our results show that BPE achieves the lowest fertility rate (1.636 at 16k vocab) by memorizing surface word forms, while achieving only 6.67% MorphAcc. PRPE achieves 83.33% MorphAcc - the highest of all systems - demonstrating that fertility rate alone is insufficient to evaluate tokenizers for agglutinative languages. All code and models are publicly available at kaggle.com/code/macmaky/quechuatok

Agentic AI development today runs on token maxing: buying capability with tokens -- longer reasoning traces, more turns, wider tool payloads, bigger replayed contexts -- so tokens per task grow faster than task value. Falling per-token prices mask the pattern; total spend rises anyway. We argue the decisive lever against token maxing is the harness: the orchestration layer that assembles context, exposes tools, sequences turns, delegates work, and carries enterprise observability and governance. We isolate it with a controlled swap: 22 locked evaluation tasks, six foundation models (Claude Sonnet 4.6, Gemini 3.1, Gemini Flash 3.5, Qwen 3.6, GLM 5.1, Palmyra X6), changing only the orchestration layer -- a frozen conventional production loop versus the Writer Agent Harness. Holding models constant, the harness cuts blended cost per task 41% (0.21->0.12), median wall-clock 44% (48s->27s), and tokens per task 38% (14.2k->8.8k), with task-completion quality at parity (0.78->0.81, directional at this sample size). Efficiency is model-invariant -- every model gets cheaper (33-61%) -- while quality gains are capability-dependent: a model's gain correlates almost perfectly with its baseline strength (r=0.99, n=6), a phenomenon we term harness leverage. Quality per dollar rises 82%; task-completions per million tokens rise from 54.9 to 92.0. On this workload the orchestration layer moved cost per task more than the full spread of the model menu did. We formalize token economics at the orchestration layer (including effective input price under prompt caching), detail the six mechanism families behind the effect -- cache-shape discipline to failure-spend governance -- compare six widely used agent systems on the same axes, and argue the harness is the one component whose efficiency multiplies across every model an organization runs -- present and future.

Algorithmic trading systems on decentralised exchanges (DEXs) reject most candidate tokens they evaluate. The counterfactual outcome of rejected candidates (what would have happened had the system entered) is rarely measured. This paper introduces Post-Rejection Follow-up Sampling (PRFS). A separate tracking subsystem samples each rejected token's price and liquidity at a configurable cadence, over a horizon of up to twenty-four hours. PRFS produces the data needed to evaluate filter precision against actual market outcomes of rejected candidates, not against synthetic backtest reconstructions. The methodology, data architecture, and deposit format are described in Section III. The companion dataset contains 67,000 forward-outcome observation rows across 2,997 rejection events spanning 457 unique mints, collected over a continuous eight-day window (2026-04-10 to 2026-04-19, UTC). Approximately 55 percent of rejection events receive at least one forward observation; coverage at the mint level is complete. The principal binding constraint on downstream classification is per-event horizon density, not event-level coverage. PRFS is dataset-independent. It generalises to any algorithmic decision system in which rejections substantially outnumber executions.

Large Language Models (LLMs) have become an indispensable part of natural language processing tasks. However, autoregressive sampling has become an efficiency bottleneck. Multi-Draft Speculative Decoding (MDSD) is a recent approach where, when generating each token, a small draft model generates multiple drafts, and the target LLM verifies them in parallel, ensuring that the final output conforms to the target model distribution. The two main design choices in MDSD are the draft sampling method and the verification algorithm. For a fixed draft sampling method, the optimal acceptance rate is a solution to an optimal transport problem, but the complexity of this problem makes it difficult to solve for the optimal acceptance rate and measure the gap between existing verification algorithms and the theoretical upper bound. This paper discusses the dual of the optimal transport problem, providing a way to efficiently compute the optimal acceptance rate. For the first time, we measure the theoretical upper bound of MDSD efficiency for vocabulary sizes in the thousands and quantify the gap between existing verification algorithms and this bound. We also compare different draft sampling methods based on their optimal acceptance rates. Our results show that the draft sampling method strongly influences the optimal acceptance rate, with sampling without replacement outperforming sampling with replacement. Additionally, existing verification algorithms do not reach the theoretical upper bound for both without replacement and with replacement sampling. Our findings suggest that carefully designed draft sampling methods can potentially improve the optimal acceptance rate and enable the development of verification algorithms that closely match the theoretical upper bound.

LLM-based Multi-Agent (LLM-MA) systems are increasingly applied to automate complex software engineering tasks such as requirements engineering, code generation, and testing. However, their operational efficiency and resource consumption remain poorly understood, hindering practical adoption due to unpredictable costs and environmental impact. To address this, we conduct an analysis of token consumption patterns in an LLM-MA system within the Software Development Life Cycle (SDLC), aiming to understand where tokens are consumed across distinct software engineering activities. We analyze execution traces from 30 software development tasks performed by the ChatDev framework using a GPT-5 reasoning model, mapping its internal phases to distinct development stages (Design, Coding, Code Completion, Code Review, Testing, and Documentation) to create a standardized evaluation framework. We then quantify and compare token distribution (input, output, reasoning) across these stages. Our preliminary findings show that the iterative Code Review stage accounts for the majority of token consumption for an average of 59.4% of tokens. Furthermore, we observe that input tokens consistently constitute the largest share of consumption for an average of 53.9%, providing empirical evidence for potentially significant inefficiencies in agentic collaboration. Our results suggest that the primary cost of agentic software engineering lies not in initial code generation but in automated refinement and verification. Our novel methodology can help practitioners predict expenses and optimize workflows, and it directs future research toward developing more token-efficient agent collaboration protocols.

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.

Block-diffusion language models are served at hand-picked operating points, such as acceptance thresholds, buffer depth, schedule, checkpoint and precision, and each point is chosen by its mean benchmark accuracy. However, a mean does not tell an operator how often a faster configuration fails on prompts that the slower one answers correctly. On the serving engine and its decode traces, the default commit rule already commits every fully resolved block, a static skip rule captures nearly all of the compute that allocation can save, and self-distillation on engine-decoded targets adds speed at unchanged accuracy. Larger speedups come from lower thresholds, which commit tokens that are still uncertain. We therefore present Redline, a finite-sample procedure that selects operating points, hand-picked or learned, from the correctness of their answers on calibration prompts. Redline keeps the reference-relative risk, the joint probability that the reference answers correctly and a candidate configuration does not, within a user-chosen budget with high probability, and deploys the fastest configuration that passes. It speeds up math at a smaller risk budget than code in both model families, and at a budget of ten percent it deploys a LLaDA2 math configuration that commits over a third more tokens in each forward. It also applies without modification to the acceptance rule of speculative decoding and to weight quantization. On the same calibration data, Redline stays within its stated failure probability, whereas each tolerance of a mean-accuracy rule either gains less speed for some model and task or exceeds the risk budget far more often for another. Code is available at https://github.com/js-lee-AI/Redline.

Per-field accept/review with selective risk at most alpha -- accept a field only if the error rate among accepted fields is controlled -- is the trust contract document-extraction systems need, and the natural procedure silently violates it on real documents. On 13,859 genuine claude-sonnet-5 fields from 800 CORD receipts (49.0% correct) we diagnose three failure modes: document clustering (design effect 1.84-2.45), score-refit leakage (coverage 0.416 at risk 0.127, violating alpha=0.10 in 95% of splits), and a tie-mass pathology (a degenerate score collapses the threshold grid, 0.030 to 0.001). We organize the fixes as a validity ladder, guarantee form stated per tier. A fit/val split protocol restores expected-selective-risk control for a learned fusion: coverage 0.318 at risk 0.096 at nominal alpha=0.10, no tolerance band (production variant 0.326) -- an on-average point whose realized risk exceeds alpha in 47.5% of resplits, not a certificate. Mondrian Learn-then-Test with exact binomial tails yields per-group PAC certificates: field-iid 0.171 at risk 0.068, cluster-corrected 0.140, doc-iid 0.060 -- the only tier matching documents, honestly near-vacuous today. Support-bin, the pre-specified provenance taxonomy, wins every rigor tier on the sonnet CORD capture (p<1e-4, Bonferroni-corrected) -- a win that does not replicate on the same documents under haiku or qwen -- while on higher-accuracy corpora pooled thresholds win: conditioning helps exactly where pooled cannot certify, subsumed by a learned score elsewhere. A frozen-configuration confirmation on selection-untouched claude-haiku-4-5 held at both risk levels, and a blind three-annotator human-gold audit verifies the practical tier's accepted-set risk at 1.3% against its 10% budget (Fleiss' kappa=0.83; labels err one-sidedly pessimistic). Released Apache-2.0 with seed-pinned, regression-gated procedures.

APort Vault is a benchmark for payment authorization in tool-using AI agents. It replays 4,371 attacks written by humans against a live payment agent during a public capture-the-flag event, across 14 models from 8 labs, five policy configurations and two replay tracks, with and without a deterministic pre-action check implementing the Open Agent Passport (OAP) specification. 225,964 evaluations completed. We report five distinct events per evaluation, because collapsing them is how an agent benchmark produces a number that does not survive review. Requests are common and their rate differs far more across configurations than across models, though each attack exists at exactly one configuration so policy and attack cohort vary together: 10.9% of model-alone evaluations at Level 1, 3.0% at Level 2, 0.1% at Level 3, 79.4% at Level 4. On the 1,293 Level 4 prompts, each evaluated on every model, request rates run from 71.2% to 84.3%, and 809 prompts (62.6%) elicited a request from all fourteen models, each ending in a successful payment to the level's allowlisted recipient. The authorization boundary is where the conditions diverge. At Levels 2 to 4, transfers to recipients the passport did not permit number 140 of 76,842 with the model alone and 0 of 69,297 behind the layer, and 105 against 0 on 68,970 matched model, prompt and track triples. The zero spans 790 source sessions, giving a per-session upper bound of 0.38%. It was not obtained by refusing payments: 25,370 payments executed behind the layer, while the policy denied 187 of the 25,640 transfer calls it evaluated, 148 of them for a forbidden recipient. We release the 225,964 evaluations, the level passports, the scoring code and the analysis script at huggingface.co/datasets/aporthq/vault-benchmark-v1 .

Long-horizon LLM agents accumulate sessions of up to millions of tokens, far more key-value (KV) cache than one GPU holds, so serving systems keep a small budget of tokens and evict the rest. Eviction methods differ in which tokens they keep but share one goal: that the tokens a later step needs are still cached. We show this is not enough. In 100K-token coding sessions on Qwen3-8B with a 2,048-token budget, a planner-guided policy that keeps over 99% of the needed tokens resident still fails late steps increasingly often: its accuracy falls from 93% to 64% as the gap between a definition and its use grows from 2K to 90K tokens, at residency 0.996. The needed tokens are resident but unretrievable: what grows is their rotary-position distance to the query. We present PACE (Planner-Aware Context Engine), a KV-cache engine that couples plan-aware retention, in which the agent's planner tells the cache which context future steps will read, with bounded positions, which renumber kept tokens so their rotary distance to the query stays below 8K at any session length. In a pre-registered evaluation on fresh tasks, PACE passes 97% of hidden unit tests at 100K tokens, against 42% for H2O and 3% for SnapKV under the same budget, and needs both components: either alone reaches 63% (the combination beats each, p < 10^-4). The result holds with length: in 1M- and 2M-token sessions on one 24 GB GPU, with definitions written 0.9M and 1.8M tokens before use, PACE passes 99% and 100% of checks and recalls every session fact, while bounded positions without foresight falls from 33% to 13%; PACE holds under 0.7 GB of KV state where the full cache would need 147-295 GB. Renumbering is exact, costs no measurable perplexity, and adds 9% wall time. On Mistral-7B, whose coding ability caps it at 40% even with nothing evicted, PACE stays at that ceiling while retention alone falls below it.

The cryptocurrency is attracting more and more attention because of the blockchain technology. Ethereum is gaining a significant popularity in blockchain community, mainly due to the fact that it is designed in a way that enables developers to write smart contracts and decentralized applications (Dapps). There are many kinds of cryptocurrency information on the social network. The risks and fraud problems behind it have pushed many countries including the United States, South Korea, and China to make warnings and set up corresponding regulations. However, the security of Ethereum smart contracts has not gained much attention. Through the Deep Learning approach, we propose a method of sentiment analysis for Ethereum's community comments. In this research, we first collected the users' cryptocurrency comments from the social network and then fed to our LSTM + CNN model for training. Then we made prediction through sentiment analysis. With our research result, we have demonstrated that both the precision and the recall of sentiment analysis can achieve 0.80+. More importantly, we deploy our sentiment analysis1 on RatingToken and Coin Master (mobile application of Cheetah Mobile Blockchain Security Center23). We can effectively provide detail information to resolve the risks of being fake and fraud problems.

Diffusion language models can write several tokens in one forward pass, and they decide during generation how many passes to spend. Two outputs of the same token length can therefore vary a lot in cost. To understand the underlying factors of this cost variation, we analyze two diffusion models across 10 tasks. We find that the number of forward passes per token depends heavily on the task. Text copied or lightly edited from the input is cheap, while open-ended writing is costly. In DiffusionGemma, generating a token for story writing requires more than three times as many forward passes as for grammar correction. Two output properties explain much of this gap: how much of it is new relative to the input, and how rare its words are within the task. The same holds within a single response when thinking is involved. The final answer is cheap when it repeats the model's reasoning trace and costly when it adds something new. In multilingual generation, we find that highly tokenized languages also require more passes per token, adding to the cost they already pay with more tokens. Finally, cost tracks correctness: on the same math problem, wrong answers take more passes per token than correct ones, and weighting candidate answers by decoding speed beats majority voting by up to 4.2 points at no extra cost. Taken together, our results show that diffusion's computational advantage depends on how much of the output is already determined, while its decoding effort offers a practical signal for making inference more efficient and reliable.

AI agents are poised to become a primary interface to digital products, acting across email, files, payments, and personal data. People without professional software backgrounds need understandable, reusable ways to control actions across services. We examine a mechanism in which a language model maps actions to plain-language consequence categories with user-authored "allow", "ask", or "never" rules. We ask what is gained and lost when decisions are made in advance as reusable rules rather than separately for each action. We analyzed 113 participants without professional software backgrounds across three conditions: per-action human-in-the-loop approval (HITL), automated per-action model review (AUTO), or user-authored consequence policy (POLICY). Participants judged 2 examples in each of 4 consequence categories; POLICY participants then set one rule per category. All supervised an 18-action simulated day, including 7 overreach actions. POLICY blocked less overreach than HITL (-20.1 percentage points, 95% CI [-32.1, -8.1]) and AUTO (-14.5 points, 95% CI [-25.8, -3.2]). POLICY lowered runtime prompts from 18.0 to 10.9, but total intervention time was not reliably lower when rule setup was included. Exploratory analysis showed that participants chose "ask" for 114 of 140 POLICY rules, returning most overreach actions to runtime. Of the 148 overreach actions executed in POLICY, 133 followed human approval and 15 ran automatically under "allow" rules. Across all 7 overreach actions, POLICY had the highest approval rate. Counterintuitively, user-authored rules did not by themselves provide stronger protection: many actions outside users' original requests went through after users approved them. These results reveal a gap between preference and commitment: repeatedly choosing "ask" preserves case-by-case choice but prevents a standing policy from settling decisions in advance.

We describe ufakzeka-1, a 151M-parameter (182M with embeddings) decoder-only Turkish language model pretrained from scratch on 13.5B tokens of openly licensed text and instruction-tuned for chat, at a total cost of about \$286 in cloud GPU, API and notebook time. The contribution is not the model's capability, which is what a model this size can be expected to have, but the record of building and measuring it: a Turkish byte-level tokenizer at 1.77 tokens per word, a three-stage pretraining schedule, a post-training mixture of openly licensed and generated data, and an evaluation battery of release gates, a rule-checked sweep of 5,508 conversations, judged conversations and hand tests, all with prompts held out from the training data, enforced by decontamination inside the data build and by a checked-in invariant script we run before each build. We report three findings that we believe transfer to other small-model efforts: a safety gate that had been "fixed" with training data written from its own questions read 64/64 while the honest figure was 34/64; training-seed variance was as large as the spread across every recipe we tried, so single-seed comparisons at this scale are uninformative; and data rounds repaired only what was absent from the data, while identity tracking over long context and multi-turn arithmetic did not move across any data change we tried, which we read as limits of the model size rather than gaps in the data, a reading the next, larger model will test. Weights, the data recipe, the evaluation code and the spend ledger are released under Apache-2.0.

Wu et al. (2026) showed that most frontier large language models (LLMs) recommend a sponsored, roughly twice-as-expensive flight when their system prompt contains a soft sponsorship cue. We reproduce their evaluation on ten open-weight chat models plus the two of their twenty-three models that are still reachable today (gpt-3.5-turbo, gpt-4o). All reported rates in this paper are produced under the same judge the original paper used (gpt-4o); we additionally store every label under an open-weight (gpt-oss-120b) and a smaller proprietary (gpt-4o-mini) judge for an ablation. Three findings emerge. First, a prose description of an LLM evaluation pipeline is not, on its own, sufficient for accurate reproduction: we surfaced three silent implementation failures that each shifted a reported rate by tens of percentage points. Second, the central claims do generalise - the gpt-3.5-turbo logistic-regression intercept of alpha = 0.81 is within four points of the original alpha = 0.86, and 200 of 200 trials on gpt-3.5-turbo and gpt-4o promote a payday lender to a financially distressed user. Third, a thirty-token user prompt that asks the assistant for a neutral comparison table first cuts sponsored recommendation from 46.9% to 1.0% averaged across our ten open-source models, and from 53.0% to 0% averaged across the two OpenAI models. AI literacy and price-comparison portals are likely market-level mitigations; the harmful-product cell is bounded by neither. Raw data, labels and analysis scripts are at https://github.com/akmaier/Paper-LLM-Ads .

Tokenization is a fundamental preprocessing step in Natural Language Processing (NLP), significantly impacting the capability of large language models (LLMs) to capture linguistic and semantic nuances. This study introduces a novel evaluation framework addressing tokenization challenges specific to morphologically-rich and low-resource languages such as Turkish. Utilizing the Turkish MMLU (TR-MMLU) dataset, comprising 6,200 multiple-choice questions from the Turkish education system, we assessed tokenizers based on vocabulary size, token count, processing time, language-specific token percentages (\%TR), and token purity (\%Pure). These newly proposed metrics measure how effectively tokenizers preserve linguistic structures. Our analysis reveals that language-specific token percentages exhibit a stronger correlation with downstream performance (e.g., MMLU scores) than token purity. Furthermore, increasing model parameters alone does not necessarily enhance linguistic performance, underscoring the importance of tailored, language-specific tokenization methods. The proposed framework establishes robust and practical tokenization standards for morphologically complex languages.

A unique aspect of ColBERT is its use of [MASK] tokens in queries to score documents (query augmentation). Prior work shows [MASK] tokens weighting non-[MASK] query terms, emphasizing certain tokens over others , rather than introducing whole new terms as initially proposed. We begin by demonstrating that a term weighting behavior previously reported for [MASK] tokens in ColBERTv1 holds for ColBERTv2. We then examine the effect of changing the number of [MASK] tokens from zero to up to four times past the query input length used in training, both for first stage retrieval, and for scoring candidates, observing an initial decrease in performance with few [MASK]s, a large increase when enough [MASK]s are added to pad queries to an average length of 32, then a plateau in performance afterwards. Additionally, we compare baseline performance to performance when the query length is extended to 128 tokens, and find that differences are small (e.g., within 1% on various metrics) and generally statistically insignificant, indicating performance does not collapse if ColBERT is presented with more [MASK] tokens than expected.

Block drafters propose several tokens in one forward pass, before earlier target tokens are realised. Their rejection mixes two losses: missing within-block path information and imperfect modelling of observable information. Accepted length cannot distinguish them. We separate the two with an information floor, the minimum expected rejection at a specified conditioning order; rejection above this floor is the model gap. Estimating both from target rollouts across four domains, four open-weight targets, and a frontier API target yields three findings. First, the all-parallel floor reaches 0.286 at the final slot on Qwen3-4B, limiting even the best proposal to 71% per-slot acceptance. Second, one realised token removes 86--100% of this floor, a locality also recovered by an independent mutual-information analysis. Third, current drafters remain far above their floors: the final-slot model gap accounts for 43--64% of DFlash rejection and 85--92% of DSpark's oracle-conditioned rejection. These findings separate the value of short-range conditioning from proposal quality. Guided by this distinction, we replace a context-independent predecessor correction with a prefix-attention head that reads committed context conditional on the predecessor. With supporting backbone components, the resulting Qwen3-4B drafter improves mean serving accepted length by 3.83% over the released DSpark checkpoint across nine tasks, without increasing the conditioning order.

The economic value of inference depends on how capacity and task information are distributed across stages of AI production. We study these organizational margins using controlled workflow experiments on externally verified software-engineering tasks. In two matched resource panels, direct execution records the same success rate of 59.6 percent at logical-token ceilings of 12,000 and 24,000, while success under information-constrained planning rises from 36.2 to 51.2 percent. The planning disadvantage narrows by 15.0 percentage points (95 percent task-cluster bootstrap interval: 4.2 to 25.8). A strict read-only planning campaign varies whether the planner sees the task issue. At 12,000 tokens, issue access raises success by about 16 percentage points over issue-hidden planning. Compared with direct execution, task-informed planning is about 10 points lower at 12,000 tokens; at 24,000 tokens, it shows a 29.6-point advantage. In the resource panels, direct execution uses substantially less than either ceiling, while the planning workflow's binding rate falls from 46.2 to 0.8 percent and downstream execution accounts for 89.9 percent of the increase in total use. Scale determines the capacity available to a system; workflow and information structure shape the productive value

Making an open-weight language model generate text is not the same as making it finish coding work. This retrospective empirical systems experience report follows my engineering work from July to October 2026 across coding clients, model-specific serving recipes, vLLM, SGLang, llama.cpp, and MLX. Models behind a nominally compatible endpoint still differed in tool-call formatting, turn control, context behavior, cache reuse, and reliability. A served model, coherent text, executed tool calls, and code that passes independent behavioral checks proved to be distinct levels of progress. The evidence includes 4,165 coding-agent gateway turns from my own sessions, 623 debug jobs, and three separate delegation studies with hidden behavioral checks and a fixed Claude Opus 5.5 parent. Saved cold-prefix requests had a median time to first token of about 69.4 s, versus 2.8 s for high-reuse requests (observational). In a 45-trial study, direct coding passed 15/15 and delegation to a quantized Qwen3-Coder 30B-A3B worker passed 12/15 and 14/15. On a second 15-task suite under a fixed artifact contract, the completed replacement cohort had 45/45 for direct coding and 12/45 for a Nemotron Cascade 2 worker (11 worker substitutions). Sixteen sessions stopped by a subscription limit before any model call were replaced; counting them as failures gives 37/45 and 8/45. A Qwen3-Coder worker passed 0/45. In both studies with a direct arm, direct coding had lower API list-price-equivalent frontier cost per verified success. A post-hoc diagnostic found that 19 of 36 artifacts rejected only for renamed parameters passed the hidden checks unchanged; they are not counted as successes. The contribution is an evidence-qualified account of serving, agent protocol, session continuity, artifact validation, and verification in practice. The studies are not pooled and do not rank models or runtimes.

Autoregressive (AR) language models generate text one token at a time, even when consecutive tokens are highly predictable given earlier context. We introduce MARS (Mask AutoRegreSsion), a lightweight fine-tuning method that teaches an instruction-tuned AR model to predict multiple tokens per forward pass. MARS adds no architectural modifications, no extra parameters, and produces a single model that can still be called exactly like the original AR model with no performance degradation. Unlike speculative decoding, which maintains a separate draft model alongside the target, or multi-head approaches such as Medusa, which attach additional prediction heads, MARS requires only continued training on existing instruction data. When generating one token per forward pass, MARS matches or exceeds the AR baseline on six standard benchmarks. When allowed to accept multiple tokens per step, it maintains baseline-level accuracy while achieving 1.5-1.7x throughput. We further develop a block-level KV caching strategy for batch inference, achieving up to 1.71x wall-clock speedup over AR with KV cache on Qwen2.5-7B. Finally, MARS supports real-time speed adjustment via confidence thresholding: under high request load, the serving system can increase throughput on the fly without swapping models or restarting, providing a practical latency-quality knob for deployment.

When a large language model (LLM) agent executes the same task, token consumption can vary by over an order of magnitude across runs. The agent chooses its next steps based on tool feedback and intermediate results, while the growing context steadily inflates the input size of every subsequent call. The total consumption of a task is therefore hard to predict before execution and the prediction must be revised as the run unfolds. In this paper, we propose TokenCast, which learns a composable cost representation for each execution segment, recording its own consumption and the context growth it introduces. Composing adjacent segments yields a cumulative estimate that captures the extra input cost incurred when context from earlier segments is re-read by every later call. As execution unfolds, newly observed evidence refreshes the forecast, requiring no additional LLM calls and incurring a mean cumulative prediction time of 32.8 ms per run on SWE-bench Verified. Across 4 task suites and 6 agent models, TokenCast's mean absolute error reduction against the strongest comparator averages 14.5% over 96 evaluated combinations. In offline budget-control replay, TokenCast uses 21.3% fewer tokens on average than a fixed-budget policy at matched trace completion. The code is available at https://github.com/DEFENSE-SEU/TokenCast.

We introduce Crypto Accounting Bench (CAB), a benchmark for assessing whether frontier and open-weight language models can reconstruct the complete journal entry that an organization actually posted for a crypto-asset transaction. CAB contains 118 evaluation tasks drawn from 7 pseudonymized organizations. Each task combines transaction mechanics, asset quantities and base-currency values, wallet and legal-entity context, counterparty evidence, related transaction legs, recurrence, tax-lot evidence, and the organization's complete chart of accounts. The target is a balanced structured entry with every required account, side, amount, currency, and full-precision asset quantity. We evaluate 12 models spanning proprietary frontier systems and open-weight releases over 3 independent attempts per task, producing 4,248 trajectories. We report 3 metrics: Mean Score, Best@3, and Pass@3. Pass@3 is the fraction of tasks with at least 1 of 3 attempts that satisfies every rubric criterion and required gate. The leading model reaches 77.43% Mean Score, while the best Pass@3 is 56.78%. Deterministic diagnostics, read from each task's best of 3 attempts and macro-averaged across the 12 models, show higher base-amount agreement (97.8%) than deciding-account accuracy (56.3%). Together with the failure analysis, these results identify account selection and complete-entry composition as the main remaining challenges on CAB.

Small language models are cheap to serve and feasible on local hardware, but strong public 135M-class systems are commonly trained with hundreds of billions to trillions of tokens on large clusters. We study a sharply resource-constrained regime: a complete 134.5M-parameter language-model pipeline executed on one NVIDIA L20 GPU. The released checkpoint, L20-Edu-135M, receives approximately 13B pretraining tokens: 10B FineWeb-Edu tokens followed by a 3B-token educational, mathematics, code, and reasoning mixture. We document the architecture, data gates, cross-source MinHash/LSH near-deduplication, segment deduplication, benchmark-overlap removal, throughput optimization, supervised fine-tuning (SFT) with weight interpolation, and reinforcement learning from verifiable rewards (RLVR) on GSM8K. In a self-run zero-shot six-task harness, L20-Edu-135M obtains a mean score of 0.4150. It trails SmolLM-135M (0.4767) and SmolLM2-135M (0.4917), but its mean is 87.1% of SmolLM-135M's while its nominal token count is 2.17% as large. This ratio is descriptive, not evidence of statistical equivalence or a controlled scaling law. The model exceeds several older 100M-160M public baselines under the same harness. Direct GRPO-style RLVR decreases GSM8K exact-match accuracy from 1.82% to 1.59% (192-token completions) and 1.21% (320-token completions). These single-run results identify a concrete failure mode rather than establishing a general lower bound on RLVR. The contribution is an auditable resource-constrained case study, not a state-of-the-art claim.

Letting a language model take gradient steps on the text it is reading, known as dynamic evaluation, lowers its loss on the rest of the document. Its behaviour on hybrid linear-attention architectures, on diffusion language models, and on standard long-context benchmarks has not been measured. We measure it. With one SGD step per 128-token chunk on the MLP output projections of the last quarter of blocks, no training, and a 2k-token window, softmax-attention models (Qwen3-1.7B, Qwen3-4B, Gemma-4-E4B) lose 5-8% of their next-token loss over a document and outperform the same frozen model given the full 32k-token context. Hybrid Gated-DeltaNet models (Qwen3.5-4B, Qwen3.8-27B) gain under 1% at every learning rate: the linear-attention blocks already perform the adaptation. The mechanism transfers to masked-diffusion models with a larger effect, 8% on Nemotron-Diffusion-8B and 17-20% on DiffusionGemma-26B, once the rate is lowered and updates are clipped. Across RULER, BABILong, LongBench-v2, SCOTUS, HotpotQA, MuSiQue, ContractNLI and citation completion, at 100 or more items on two model sizes, weights updated on the full document recover none of the accuracy lost to context truncation: the adaptation is style and vocabulary, not retrieval. Writing facts with the next-token loss stores the prompt format rather than the fact (30% recall under chat prompting against 80% in context) and costs 36-38 GSM8K points at every dose. Writing a conclusion as behaviour instead, by distilling the model's own conclusion-conditioned judgments into a low-rank adapter, reaches the in-context ceiling (0.83 and 0.88 against 0.84 and 0.88 on 100 games) with GSM8K unchanged, and the conclusion can come from an inquiry loop that found it by asking information-gain questions.

Keyword-matching benchmarks can credit small models for tool use they never perform. We document such a false positive in a matched-architecture pair of Spanish security language models and propose a ladder of strict, cheap diagnostics. A 661.6M parameter model (approx. 65% code/technical text; no dedicated SFT) and a 1,109M model (web-heavy multi-phase curriculum; 6B-token tool-SFT) share decoder, tokenizer, and special tokens, scoring almost identically on lenient tool-use metrics (B4: 0.660 vs. 0.650). Verbatim-reproduction checks on training examples separate them completely: the 600M emits valid tool calls with generalized arguments on 6/6 examples; the 1B does so on 0/6 across checkpoints. A first-token probe localizes the 1B's failure to a missing prior (prob. 10^{-4}--10^{-5} on <|tool_call|>), which was erased by its web-heavy training phase. A targeted SFT recipe (diverse corpus, 5x higher learning rate, 2,202 steps, ~3.3 GPU-hours) repairs the 1B using three orders of magnitude fewer tokens than the failed phase. On all 269 corpus rows, valid emission rises from 0.100 to 0.959 (600M: 0.926). On 238 unseen prompts, the repaired 1B passes 0.536 vs. the 600M's 0.428 (p = 0.004). Embedding-drift checks show the repair did not move the trigger token's tied embedding (97.7% of the bf16 table remains bit-identical), meaning changes live in the surrounding network. Both models over-trigger, rarely answering negative prompts without a call (0.09 for 600M, 0.17 for repaired 1B). Factorial analyses confirm all repair configurations install the format, though suppression benefits from a diverse corpus remain a hypothesis due to seed sensitivity. This cheap diagnostic ladder costs minutes of CPU time and should gate tool-use claims on small models.

Speculative decoding accelerates LLM inference by drafting future tokens with a small model, but drafter models degrade sharply under template perturbation and long-context inputs. We identify a previously-unreported phenomenon we call attention drift: as the drafter generates successive tokens within a speculation chain, attention progressively moves from the prompt onto its own recently-generated tokens. We observe this across both EAGLE3 drafters and MTP heads, suggesting drift is a property of drafter designs. We trace this to the un-normalized residual path between chain steps: the drafter's hidden state magnitude grows monotonically with chain depth, which exhibits dynamics consistent with additional pre-norm transformer layers stacked on the target rather than as a standalone autoregressive predictor. In order to limit the growth, we propose two architectural changes: Post-norm on the drafter hidden states and per-hidden-state RMSNorm after capturing target hidden states. Our interventions improve acceptance length over the current leading model, pre-norm EAGLE3, by up to 2times under template perturbation, 1.18times on long-context tasks, and 1.10times on seven standard benchmarks spanning multi-turn chat, math, and coding. Our changes also allow shorter train-time-test depths to generalize over longer drafting sequences.

Intent-based Open RAN needs an interpreter that turns intents into A1 policies within the loop of the RAN intelligent controller (RIC). Decision models such as Jev-1.13.0 return typed policy fields, whereas generative large language models (LLMs) produce the policy token by token. We ask whether the extra delay of LLMs costs control deadlines, RIC capacity, or radio performance. We compare Jev-1.13.0 and two other decision models with LLMs on the RANIntent v1 benchmark, in closed-loop ns-3 simulation and on a real A1 and E2 path. Median interpretation takes 0.286 to 2.35 s, against under 25 ms for A1 and E2 transfer. Jev-1.13.0 meets the 1 s near-real-time budget on 99.8% of calls, while two hosted LLMs meet it on 17.9% and 0%. In the radio network, ideal enforcement moves the affected-class service-level agreement (SLA) violation by 3.96 percentage points in the direction each intent requests, against no update at the base point. No hosted LLM showed a resolved increase over Jev-1.13.0 at that point. At the same point, per-second direct control gave no resolved SLA reduction over a numerical xApp. Slow interpreters miss the 1 s budget, and two interpreters saturate their queues at 2 intents/s, whereas no radio penalty of slow interpreters was resolved at the base point.

Scientific agents increasingly analyze data, execute code, and produce research artifacts, yet most benchmarks emphasize final answers, isolated programs, or a single domain. We introduce FrontierChallenge, a cross-domain benchmark comprising 300 end-to-end scientific workflows. In this paper, we release and evaluate 97 of these tasks, spanning quantum chemistry, molecular dynamics, materials characterization, analytical chemistry, life science, and electrochemistry/environment. Each task provides fixed inputs and specifies a bundle of required scientific deliverables. We evaluate twelve frontier models with three agent scaffolds. Pass Rate measures the fraction of tasks satisfying the full-completion criterion, while Avg. Score captures partial progress. Each of the best-performing configurations completed only 20 of the 97 released tasks, yielding a Pass Rate of 20.6%. Partial progress translated especially poorly into complete delivery in analytical chemistry and electrochemistry/environment: Avg. Scores reached 87.6 and 94.9, but the highest Pass Rates were only 4% and 0%. Among non-passing Claude Code trajectories, 75.5% still ended with language claiming completion. These findings show that neither high partial scores nor confident claims of completion reliably indicate that a scientific task has been fully delivered, highlighting the need to evaluate end-to-end workflow execution and the completeness of scientific deliverables together.

Token compression aims to speed up large-scale vision transformers (e.g. ViTs) by pruning (dropping) or merging tokens. It is an important but challenging task. Although recent advanced approaches achieved great success, they need to carefully handcraft a compression rate (i.e. number of tokens to remove), which is tedious and leads to sub-optimal performance. To tackle this problem, we propose Differentiable Compression Rate (DiffRate), a novel token compression method that has several appealing properties prior arts do not have. First, DiffRate enables propagating the loss function's gradient onto the compression ratio, which is considered as a non-differentiable hyperparameter in previous work. In this case, different layers can automatically learn different compression rates layer-wisely without extra overhead. Second, token pruning and merging can be naturally performed simultaneously in DiffRate, while they were isolated in previous works. Third, extensive experiments demonstrate that DiffRate achieves state-of-the-art performance. For example, by applying the learned layer-wise compression rates to an off-the-shelf ViT-H (MAE) model, we achieve a 40% FLOPs reduction and a 1.5x throughput improvement, with a minor accuracy drop of 0.16% on ImageNet without fine-tuning, even outperforming previous methods with fine-tuning. Codes and models are available at https://github.com/OpenGVLab/DiffRate.

Foundation models tokenize Ukrainian legal text with vastly different efficiency, yet no systematic comparison exists for this domain. We benchmark seven models from five providers on 273 validated court decisions from Ukraine's state registry (EDRSR), measuring tokenizer fertility and zero-shot performance on three tasks. Three findings emerge. (1) Tokenizer fertility varies 1.6x: Qwen3 models consume 60% more tokens than Llama-family models on identical input, directly reducing API cost. (2) NVIDIA Nemotron Super 3 (120B) achieves the highest composite score (83.1), outperforming Mistral Large 3 (675B total, 41B active) -- a model with 5.6x more total parameters and 3.4x more active parameters per token -- at one-third the API cost. (3) Few-shot prompting degrades performance by up to 26 percentage points; stratified and prompt-sensitivity ablations confirm this is intrinsic to Ukrainian-language demonstrations, not an artifact of example selection. For practitioners: tokenizer analysis should precede model selection, and zero-shot is a more reliable default than few-shot for morphologically rich languages.

Speculative decoding typically characterizes speedup by draft tokens accepted per verification round. Lossless verification preserves the target model's output distribution and thus expected completion length, making this metric a meaningful efficiency proxy. Relaxed (lossy) verification accepts tokens that lossless verification would reject, potentially altering that distribution. We identify a length inflation phenomenon: relaxed verification lengthens completions, primarily by extending reasoning traces, largely offsetting apparent gains in accepted tokens per round. We evaluate existing training-free relaxed verification rules on twelve target-drafter pairs against lossless verification on identical prompts, measuring completion length, verifier rounds, wall-clock time, and task accuracy. At each rule's most relaxed evaluated setting, relaxed verification on average accepts 1.34times as many tokens per round but produces 1.54times longer completions. For rules with higher acceptance lengths, tokens accepted only under relaxed verification tend to have low target-model probability, contributing to length inflation. For rules that reduce verifier rounds, accuracy typically drops substantially. Further analysis traces some failures to continued reasoning that exhausts the budget needed to deliver a final answer. Overall, accepted tokens per round alone is an unreliable measure of decoding speed under relaxed verification. Completion length and verifier rounds must also be reported to assess actual efficiency.

Hierarchical two-hop dispatch can reduce slow-fabric traffic in expert-parallel Mixture-of-Experts training, but it adds a second collective and an arrival-side operator chain. We present a cost model for screening that trade at the communication-call level, bounded by validation gates that withdraw a capability in code when they fail rather than reporting a caveat. At a reference geometry with 16 groups of 8 ranks, q=3, H=2048, and 4096 tokens per rank, the corrected effective breakeven hierarchy ratio is 3.98 for the measured PyTorch arrival chain, 1.49 for a hypothetical fused target, and 1.10 at zero implementation overhead. These are ratio-only sensitivity results, not deployment predictions: platform A measures 1.03, platform B has no separated fast/slow measurement, and neither machine measured here reaches the hierarchical regime. Four communication-level corpora pass their gates; a drift probe and the step-level gate fail. The latter failure is enforced in code, so we make no training-throughput prediction. The enabling routing constraint fixes per-token fan-out and per-selected-group quota, while aggregate per-peer counts remain data-dependent. Its measured validation-loss cost is small but nonzero (+0.0034 nats); downstream equivalence is reported with incomplete estimator provenance and is therefore not independently reconstructible from the artifact. Code, calibration constants and the validation gates are at https://github.com/weich97/TerraceMoE-simulator.

The problem of pre-training data detection for large language models (LLMs) has received growing attention due to its implications in critical issues like copyright violation and test data contamination. The current state-of-the-art approach, Min-K%, measures the raw token probability which we argue may not be the most informative signal. Instead, we propose Min-K%++ to normalize the token probability with statistics of the categorical distribution over the whole vocabulary, which accurately reflects the relative likelihood of the target token compared with other candidate tokens in the vocabulary. Theoretically, we back up our method by showing that the statistic it estimates is explicitly optimized during LLM training, thus serving as a reliable indicator for detecting training data. Empirically, on the WikiMIA benchmark, Min-K%++ outperforms the SOTA Min-K% by 6.2% to 10.5% in detection AUROC averaged over five models. On the more challenging MIMIR benchmark, Min-K%++ consistently improves upon Min-K% and performs on par with reference-based method, despite not requiring an extra reference model.

The emergence of Large Language Models (LLMs) has significantly influenced various aspects of software development activities. Despite their benefits, LLMs also pose notable risks, including the potential to generate harmful content and being abused by malicious developers to create malicious code. Several previous studies have focused on the ability of LLMs to resist the generation of harmful content that violates human ethical standards, such as biased or offensive content. However, there is no research evaluating the ability of LLMs to resist malicious code generation. To fill this gap, we propose RMCBench, the first benchmark comprising 473 prompts designed to assess the ability of LLMs to resist malicious code generation. This benchmark employs two scenarios: a text-to-code scenario, where LLMs are prompted with descriptions to generate code, and a code-to-code scenario, where LLMs translate or complete existing malicious code. Based on RMCBench, we conduct an empirical study on 11 representative LLMs to assess their ability to resist malicious code generation. Our findings indicate that current LLMs have a limited ability to resist malicious code generation with an average refusal rate of 40.36% in text-to-code scenario and 11.52% in code-to-code scenario. The average refusal rate of all LLMs in RMCBench is only 28.71%; ChatGPT-4 has a refusal rate of only 35.73%. We also analyze the factors that affect LLMs' ability to resist malicious code generation and provide implications for developers to enhance model robustness.

A key component of building safe and reliable language models is enabling the models to appropriately refuse to follow certain instructions or answer certain questions. We may want models to output refusal messages for various categories of user queries, for example, ill-posed questions, instructions for committing illegal acts, or queries which require information past the model's knowledge horizon. Engineering models that refuse to answer such questions is complicated by the fact that an individual may want their model to exhibit varying levels of sensitivity for refusing queries of various categories, and different users may want different refusal rates. The current default approach involves training multiple models with varying proportions of refusal messages from each category to achieve the desired refusal rates, which is computationally expensive and may require training a new model to accommodate each user's desired preference over refusal rates. To address these challenges, we propose refusal tokens, one such token for each refusal category or a single refusal token, which are prepended to the model's responses during training. We then show how to increase or decrease the probability of generating the refusal token for each category during inference to steer the model's refusal behavior. Refusal tokens enable controlling a single model's refusal rates without the need of any further fine-tuning, but only by selectively intervening during generation.

RED-2400 is a public benchmark of algorithmically-rejected trading events from a live Solana decentralized-exchange filter stack. I logged the data continuously between 2026-04-10 and 2026-05-02. The benchmark contains 6,659 rejection events linked to 169,122 post-rejection price and liquidity observations and 1,836 graveyard-tracker snapshots. Outcome labels follow the five-tier classification of Kamat (2026c): saved (windowed), saved (early-death), missed, flat, and unclassifiable. Thresholds use the trough-to-reference and peak-to-reference price ratios within a 24-hour window. Most filter-design datasets cover the accept side only. That gap leaves reject-side outcomes unmeasured and biases filter validation. RED-2400 lets researchers replicate filter-precision claims directly. RED-2400 is the first window in a planned dataset series; subsequent windows will extend the time horizon and enable regime-stratified analysis.

DeFi investment agents, systems that use AI for autonomous on-chain trading, have attained over USD 3 billion in combined token valuations since late 2024. We survey over 1,900 AI-tagged crypto projects, filter to investment-focused agents, and curate 10 representative projects spanning strategy and observability dimensions. We then conduct a deep-dive architectural analysis of two prominent agent frameworks, ElizaOS and Virtuals Protocol, and a quantitative on-chain performance analysis of 11 Solana-based agent treasuries with publicly attributable trading activity, covering 925,323 token holders. We find that current deployments remain early and heterogeneous: (1) in our sample, many projects do not yet provide clear evidence of autonomous trade execution, and developer interviews suggest that many visible deployments remain basic API integrations; (2) agent treasuries retain over USD 30M in paper gains while token holders collectively lost USD 191.7M, with the top 1% of wallets capturing 81.4% of all gains (USD 1.81B); (3) token valuations are weakly connected to treasury fundamentals, with market-cap-to-AUM ratios exceeding 10,000x versus below 1x for established DeFi protocols; and (4) aggregate user gains peaked at USD 2.4B before declining to net losses, with median returns negative on every platform and tokens declining 93% on average from all-time highs. We interpret these outcomes as characteristic of a permissionless, first-generation market in which open infrastructure enables rapid experimentation but also allows naive or speculative agents to launch before robust standards for autonomy, performance, and stakeholder alignment emerge. We therefore propose a maturity framework along three dimensions: autonomous execution, risk-adjusted profitability, and stakeholder alignment, to characterize the gap between current deployments and future investment-grade agent systems.

Split language models keep input text on a client but send intermediate representations to a server. Attacks can reconstruct text from these messages. Comparing privacy evaluations is difficult when attackers see different messages, know different model weights or use different datasets and recovery measures. We present LLM-Attacker, a framework that records these choices and lets researchers check, compare and extend an evaluation. Compatible attacks reuse recorded messages, and their predictions are saved for independent scoring. The scoring rule requires a linked undefended reference before a defence can pass. Reports also identify attacks that remain untested. In a controlled GPT-2 experiment, adding gradients returned during training increased token recovery from 94.20% to 97.38% and exact-document recovery from 13.71% to 37.77%. The 95% interval for the token gain lay above the two-percentage-point threshold fixed before testing. This experiment is shared with a companion study; here, it tests how the evaluation responds to a change in attacker access. Further studies measure how corpus choice and compression defences affect recovery, compare an adapted published attack with its reference implementation, and test the measurement code. The framework reports implementation checks separately from measured attack effectiveness. Each privacy claim should state the attacker, dataset and scoring rule, report its controls and identify untested attacks.

The increasing demand for large language model (LLM) serving has necessitated significant advancements in the optimization and profiling of LLM inference systems. As these models become integral to a wide range of applications, the need for efficient and scalable serving solutions has grown exponentially. This work introduces TokenSim, a comprehensive hardware and software exploration system designed specifically for LLM inference. TokenSim is characterized by its support for extensible system optimizations including scheduling and memory management. We validate the results with systems running with realworld datasets, achieving an error rate of less than 1%. Furthermore, TokenSim facilitates various insightful explorations into the performance and optimization of LLM serving systems.

The number of tokens it takes to encode parallel text in different languages is known to vary. These disparities are called token premiums. Having high token premiums leads to less throughput during training and increases costs at inference. In this paper, we show that even after controlling for dataset size, vocabulary size, and data content, monolingual tokenizers exhibit a wide range of token premiums across languages. To understand the cross-linguistic differences that cause these token premiums, we train a suite of approximately 7,000 comparable monolingual tokenizers for 97 languages, manipulating tokenization algorithm, vocabulary size, and dataset size. We measure token premiums and test for a relationship between factors such as data similarity (between tokenizer training and evaluation), vocabulary size, and pre-tokenization. We also investigate the role of language-specific features such as writing system and word length. We find that similarity between training and test data does not impact token premiums, but vocabulary size and pre-tokenization do. While simply increasing vocabulary size does not lead to reduced token premium effects, we can determine an ``optimal'' vocabulary size for each language to achieve significantly reduced token premium effects. We also train superword tokenizers which allow merges over whitespaces, and we find that they both reduce token premium effects and improve compression overall. Thus, intervening on the vocabulary size or the pre-tokenizer significantly reduces crosslingual token premium effects.

Brazilian Portuguese remains under-served by open language models, and the few that exist are difficult to reproduce and are often compared without measures of uncertainty. We release Manacá-1B, an open decoder-only model of 1.72 billion parameters trained from scratch for Brazilian Portuguese with a fully containerized, reproducible pipeline. The pretraining is stable, with zero skipped or NaN steps and self-recovering loss spikes, and we release its full log and dynamics. We evaluate the model against nine open baselines on four Portuguese benchmarks under a single harness. Every comparison reports a standard error and a paired significance test, and the harness is validated against previously published numbers. On last-word prediction Manacá-1B is the strongest model below the 7B scale, exceeding both Tucano-1b1 and Tucano-2b4 on LAMBADA-PT with large paired margins; it is competitive on commonsense completion and near chance on multiple-choice reasoning, as are all small base models. Along the way we document a concrete evaluation pitfall: converting a SentencePiece tokenizer with case-folding normalization to the HuggingFace fast format silently drops the normalizer, routing every capitalized token to byte-fallback and depressing scores in a way that is invisible in aggregate metrics. The uncorrected tokenizer lowered LAMBADA-PT accuracy from 45.3 to 25.0; we quantify the effect and provide a one-line fix that reproduces the training tokenizer exactly. Code, raw training and evaluation logs, per-example prediction vectors, the model weights, and the corrected tokenizer are released so that every number in this paper can be recomputed.

Practitioners make three prompt-design decisions with almost no controlled evidence behind them: how to format instructions and context (markdown, plain text, prose, or tabular), how many simultaneous instructions a system prompt can carry before compliance degrades, and how much context a model can hold before recall and honesty degrade. We report two controlled experiments crossing all three factors on one held, contamination-free synthetic corpus (the "Book of Veyra," 8,780 uniquely-named entities, deterministically regenerable from a fixed seed), evaluated across five models. Experiment 1 (960 calls/model) measures instruction-following decay as rule count N grows from 10 to 160, crossed with four formats and system-prompt vs. user-turn placement. Perfect-response rate collapses to zero by N=80 for every model, format, and placement. Placement produces effects at least as large as format at N=160 in most models, but the direction is model-specific. No model shows a reliable markdown advantage; one 35B model favors plain text instead. Experiment 2 (5,520 calls/model) measures recall accuracy, false-premise sycophancy, and absent-fact fabrication across a 2k-to-512k-token context ladder in the same four formats. Recall stays near ceiling through 64-128k tokens, then degrades sharply and format-dependently: one model's accuracy spread reaches 48 points at 128k tokens. Fabrication never occurs (0/5,760 probes), and sycophancy stays negligible (<=8.3%). What rises sharply near each model's context ceiling is outright refusal to answer (0% to 79-90%), distinct from sycophancy or fabrication. Neither pre-registered format ordering holds, and token overhead (+22% to +37% over plain text) further changes which format is preferable where accuracy spread is genuine. We release the full harness, corpus generator, and raw results (VeyraBench): https://github.com/iNetanel/veyrabench

Speculative decoding accelerates autoregressive LLMs by having a lightweight drafter propose tokens that the target model verifies in parallel. Diffusion-based drafters further reduce drafting latency by proposing multiple tokens at once. However, these tokens are predicted independently, so a single early error causes prefix verification to discard the rest of the draft, even when it contains useful downstream predictions. We introduce DEdit, a diffusion-based drafter that can not only draft by conventional parallel unmasking but also iteratively edit its draft through token-to-token predictions. Through editing, later predictions can serve as bidirectional context for repairing earlier errors and extending the accepted prefix. To teach the model to repair errors while preserving correct predictions, we propose ProposalMix, a training scheme that mixes draft predictions with ground-truth tokens based on first-pass confidence during training. Across seven benchmarks on Qwen3-4B and Qwen3-8B, DEdit achieves the highest macro-average token acceptance and speedup among the evaluated drafters, reaching macro-average speedups of 5.72times and 5.97times over autoregressive generation under greedy decoding, respectively. Further analysis shows that acceptance improves with more editing passes and wider drafting windows, and that ProposalMix halves harmful edits that shorten the accepted prefix. Moreover, restricting the editor to causal attention lowers acceptance, especially on highly predictable outputs, indicating that future context is a key source of these gains.

With the massive surge in ML models on platforms like Hugging Face, users often lose track and struggle to choose the best model for their downstream tasks, frequently relying on model popularity indicated by download counts, likes, or recency. We investigate whether this popularity aligns with actual model performance and how the comprehensiveness of model documentation correlates with both popularity and performance. In our study, we evaluated a comprehensive set of 500 Sentiment Analysis models on Hugging Face. This evaluation involved massive annotation efforts, with human annotators completing nearly 80,000 annotations, alongside extensive model training and evaluation. Our findings reveal that model popularity does not necessarily correlate with performance. Additionally, we identify critical inconsistencies in model card reporting: approximately 80\% of the models analyzed lack detailed information about the model, training, and evaluation processes. Furthermore, about 88\% of model authors overstate their models' performance in the model cards. Based on our findings, we provide a checklist of guidelines for users to choose good models for downstream tasks.

User representation learning in real-world industrial scenarios is commonly scaled by increasing user amount, behavioral sequence length and model size. However, existing methods face two challenges: (i) Bottleneck for raw data scaling at billion-scale capacity, as performance exhibit diminishing performance gains with larger-scale raw text user behavioral input, which can be mitigated by tokenization. (ii) Lack of quantitative analysis of how tokenization configurations should scale with data size. In this report, we propose User Behavioral Densing Law for characterizing the quantitative relationship between data scale and the minimum sufficient tokenization capacity. Firstly, we conduct a pilot study on raw & tokenized scaling comparison on billion-scale Alipay dataset, revealing the raw data scaling bottleneck and the sustained gains enabled by tokenization. To derive the scaling pattern governing the minimum sufficient tokenization configuration at different data scales, theoretical analysis and systematic experiments are employed to summarize the quantitative scaling pattern. We find an approximately linear relationship between the logarithms of minimum sufficient tokenization capacity and input data size measured by tokens, and the scaling slope varies systematically with the tokenization method and data source, reflecting differences in representation-space redundancy and intra-source uniqueness. Guided by the proposed law, we further develop ALGN, an adaptive variable-length tokenization method that improves capacity allocation. Extensive experiments across diverse data sources, tokenization methods, and downstream tasks demonstrate the generalizability and reliability of the User Behavioral Densing Law, providing practical guidance for tokenization configuration selection in large-scale user representation learning. Moreover, ALGN outperforms existing baselines.

AI agents may consume massive tokens across a wide range of tasks, but how many of those tokens are really necessary? Although prior work has shown that agents can incur unexpectedly high token costs, we still lack a systematic way to identify and measure token waste. In this work, we introduce a unified framework for measuring token waste across heterogeneous agent systems. Our framework records and normalizes complete agent trajectories, identifies removable spans using an LLM-as-a-Judge (LLMaJ), and maps the identified text spans back to the original trajectory for token-level measurement. We validate the identified waste through direct interventions by removing marked spans, resuming agent execution, and measuring changes in task outcomes and subsequent actions. In online pruning experiments, we remove 41-46% of the model's reasoning while largely preserving task performance across two distinct agent benchmarks. Using this framework, we analyze agents across multiple settings. Our analysis shows that token waste has a system-level structure that aggregate token usage alone cannot reveal. First, waste varies more across tasks and agent settings than across models. Second, reasoning and visible execution can exhibit different waste patterns. Third, different harnesses can introduce distinct sources of inefficiency even under the same model and task. Overall, our framework transforms token waste into an actionable optimization signal, enabling researchers and practitioners to train more efficient models, design better reasoning strategies, and optimize agent harnesses toward reducing unnecessary computation rather than merely reducing token usage.

Tokenizers play a crucial role in determining the performance, training efficiency, and the inference cost of Large Language Models (LLMs). Designing effective tokenizers for multilingual LLMs is particularly challenging due to diverse scripts and rich morphological variation. While subword methods like Byte Pair Encoding (BPE) are widely adopted, their effectiveness in multilingual settings remains underexplored. We present MUTANT, a recipe for building multilingual tokenizers, with careful vocabulary and training data design, language-aware pre-tokenization, and subword and multiword aware training. We also introduce MUTANT-Indic, a tokenizer for India-specific multilingual LLMs, that produces linguistically coherent tokens and achieves state-of-the-art performance. Evaluated across English, 22 Indian languages and code data, our tokenizer improves the average fertility score by 39.5%$ over LLaMA4 and by 18% over Sutra (the current best). This translates to 44% improvement in inference throughput over LLaMA4 while maintaining comparable performance on English and Indic benchmarks. We present detailed ablations across tokenizer training data size, vocabulary size, merging techniques, and pre-tokenization strategies, demonstrating the robustness of our design choices.

Byte-level BPE tokenizers that use the HuggingFace ByteLevel pre-tokenizer inherit GPT-2's word regex, where a word is defined as L+, one or more Unicode letters. In abugida scripts, vowels are written as combining marks; this pattern therefore splits each word at every vowel sign. Since BPE merges only within a pre-token, those splits persist through training regardless of vocabulary size or corpus composition. We formalise this effect as a training-free lower bound on fertility. Across 26 languages from a parallel corpus, every one of the 17 abugidas is affected, ranging from 1.47x (Tibetan) to 9.02x (Thai), whereas Latin, Cyrillic, Hangul, and Han show exactly 1.00x. For 5 languages, matched tokenizer pairs that differ only in this character class fall within 2.2% of the predicted floor, scoring 4.78 versus 1.58 tokens per word on Nepali. When the Nepali share of the training corpus is swept from 5% to 95%, the broken tokenizer barely shifts at all (1.7%) while the fixed one shifts 33.9%, which separates a structural ceiling from a data shortage without needing to inspect any code. We train three 268M models that differ only in their tokenizer; the fixed variant achieves 4.43% lower held-out Nepali bits per byte at equal compute, and it still leads when given the same bytes with 1.59x the compute. A census of 3,479 HuggingFace repositories finds the letters-only word class present in 63.3% of the most-downloaded text-generation models, accounting for 72.5% of their downloads. GPT-4o's o200k pattern already uses a mark-aware word class, making the repair itself prior art. We quantify its value, show how to recognise its absence from symptoms alone, map which scripts it reaches, measure how widely it is deployed, and release a 65,536-entry Nepali-English tokenizer with a harness that regenerates every number here from public data on a laptop.

Evaluation budgets in agentic retrieval-augmented generation span questions, search trajectories, and repeated answers. We measure allocation precision, reading efficiency, and cost boundaries using a retrieval-feedback comparison on HotpotQA and MuSiQue. At 34.14--34.39M model tokens, broader question coverage lowers standard error by 33\% versus five reads and 12.6\% versus three trajectories. Archived nested and Q-only forecasts predict these allocations within 4.0\% and 3.5\%, respectively. Depth subsets establish no clear forecasting advantage beyond the two-trajectory audit. One-read variance penalties relative to the fitted optimum at the same token budget are 0--9.9\%, with substantial Pro uncertainty. Under recorded model fees, more questions beat more trajectories at search prices of \$0--1 per 1,000 requests; question-versus-read fee rankings remain unresolved. Temperature zero cuts answer disagreement from 14.3\% to 3.4\% while comparison precision stays similar. \par\medskip\noindentKeywords: Agentic RAG; Evaluation budget; Generalizability theory; Repeated sampling.

Many sentiment datasets use one annotator for each item, so the test labels show the conventions of that person. We build a Nepali benchmark from two annotators who labelled the same sentences blind. They labelled 5,614 sentences from tweets of Nepali public figures on a five-point scale. We keep only the 3,189 sentences where both give the same polarity. On the three-class benchmark, NepBERTa with a fixed tokenizer gets a macro-F1 of 0.877 and IndicBERTv2 0.867. Three commercial large language models (LLMs) get 0.882 to 0.896 zero-shot, and these differences are not significant. On the five-point scale, 766 sentences have two adjacent labels, for example Pos and SPos. We accept both labels and train on both. On this task, Claude Opus 5.5 beats the best encoders by 0.10 to 0.12 macro-F1. A tokenizer audit adds a second result. The default tokenizers of three Nepali BERT models delete the non-spacing Devanagari vowel signs and the virama. When we restore the marks, macro-F1 increases by 0.084 for NepBERTa but decreases by 0.095 for a model pretrained on stripped text.

Solving complex or long-horizon problems often requires large language models (LLMs) to use external tools and operate over a significantly longer context window. New LLMs enable longer context windows and support tool calling capabilities. Prior works have focused mainly on evaluation of LLMs on long-context prompts, leaving agentic setup relatively unexplored, both from capability and safety perspectives. Our work addresses this gap. We find that LLM agents could be sensitive to length, type, and placement of the context, exhibiting unexpected and inconsistent shifts in task performance and in refusals to execute harmful requests. Models with 1M-2M token context windows show severe degradation already at 100K tokens, with performance drops exceeding 50\% for both benign and harmful tasks. Refusal rates shift unpredictably: GPT-4.1-nano increases from sim5\% to sim40\% while Grok 4 Fast decreases from sim80\% to sim10\% at 200K tokens. Our work shows potential safety issues with agents operating on longer context and opens additional questions on the current metrics and paradigm for evaluating LLM agent safety on long multi-step tasks. In particular, our results on LLM agents reveal a notable divergence in both capability and safety performance compared to prior evaluations of LLMs on similar criteria.

Mixture-of-experts models make nearly trillion-parameter capacity accessible with sparse per-token computation, provided that the serving system can distribute the weights and coordinate their execution. We present Cascadia's resident execution of Inkling, a 975B-total/41B-active-parameter model, on eleven Intel Core Ultra X7 358H AI PCs, each with 64 GB of memory, Arc B390 integrated graphics and gigabit Ethernet. We contribute a custom resident MoE engine that preserves Inkling's routing rules, constructs compressed graphs for OpenVINO's fused iGPU primitives, and coordinates FP16 expert computation with FP32 output restoration. The engine fits six consecutive decoder layers per machine and represents dense feed-forward blocks as all-active expert slices, reducing measured dense-layer call time from approximately 8.1 to 4.5 ms. A streaming pipeline coordinates concurrent generation, while captured-state draft evaluation measures agreement with the deployed numerical path. Paired measurements at fifteen concurrency levels from 1 to 176 streams reach 60.29 aggregate decode tokens/s at 88 streams, with 46.87 tokens/s over the complete serving phases. At fifteen streams, median first-token latency is 6.05 s. Raising the context budget from the 1,024-position default, real prompts of 1k to 64k tokens recover the embedded code in all 19 measured answers, with first-token time growing as aN+bN^2 and decode latency growing approximately linearly, both bounded by a single-threaded CPU attention loop rather than by memory, which holds 512k positions per stream. Evaluation on captured fleet states separates the effects of vocabulary selection and weight quantization on draft agreement. Together, these contributions establish an execution and evaluation approach for large sparse models on distributed client systems with shared CPU-GPU memory.

Training of a tokenizer plays an important role in the performance of deep learning models. This research aims to understand the performance of tokenizers in five state-of-the-art (SOTA) large language models (LLMs) in the Assamese language of India. The research is important to understand the multi-lingual support for a low-resourced language such as Assamese. Our research reveals that the tokenizer of SUTRA from Two AI performs the best with an average Normalized Sequence Length (NSL) value of 0.45, closely followed by the tokenizer of GPT-4o from Open AI with an average NSL value of 0.54, followed by Gemma 2, Meta Llama 3.1, and Mistral Large Instruct 2407 with an average NSL value of 0.82, 1.4, and 1.48 respectively.

Automatic speech recognition (ASR) systems are commonly evaluated with word error rate (WER), yet many voice workflows depend on exact written values for identifiers, paths, and measured quantities. A transcript can appear fluent and achieve low WER while corrupting a value that a downstream system must parse, store, or execute. We introduce VoiceCodeBench, a benchmark for evaluating exact structured-token recovery in English ASR. It contains 300 human-recorded workplace segments spanning eight workflow domains and 1,482 audited target entities across 26 entity types, each with a canonical written form recoverable from the audio. Under a raw-audio-only protocol, systems receive audio bytes without additional context or metadata. Alongside WER, we evaluate Canonical Token/Entity Match (CTEM), Task Success Rate (TSR), and per-type exact recovery. Across 12 baseline ASR systems, lower WER generally corresponded to better structured-token recovery but did not fully determine it: Spearman correlations were -0.73 for both WER versus CTEM and WER versus TSR. The strongest baseline by TSR reached only 68.7%, leaving nearly one third of recordings with at least one unrecovered workflow-critical value. These results show that entity-sensitive metrics are needed to assess whether ASR output preserves exact values that production systems must parse, route, store, compare, or execute.

Scaling laws are useful guides for developing language models, but there are still gaps between current scaling studies and how language models are ultimately trained and evaluated. For instance, scaling is usually studied in the compute-optimal training regime (i.e., "Chinchilla optimal" regime); however, in practice, models are often over-trained to reduce inference costs. Moreover, scaling laws mostly predict loss on next-token prediction, but ultimately models are compared based on downstream task performance. In this paper, we address both shortcomings. To do so, we create a testbed of 104 models with 0.011B to 6.9B parameters trained with various numbers of tokens on three data distributions. First, we investigate scaling in the over-trained regime. We fit scaling laws that extrapolate in both the number of model parameters and the ratio of training tokens to parameters. This enables us to predict the validation loss of a 1.4B parameter, 900B token run (i.e., 32times over-trained) and a 6.9B parameter, 138B token runx2014each from experiments that take 300times less compute. Second, we relate the perplexity of a language model to its downstream task performance via a power law. We use this law to predict top-1 error averaged over downstream tasks for the two aforementioned models using experiments that take 20times less compute. Our experiments are available at https://github.com/mlfoundations/scaling.

Annotation quality is difficult to sustain when campaigns span weeks or months with small annotator pools. We present a Setswana sentiment dataset of 3,565 tweets annotated by three native-speaker annotators across eight batches and examine why inter-annotator agreement (IAA) declines over time. Despite an aggregate Randolph's free-marginal Kappa of κ= 0.76, "excellent," per-batch κ falls by more than 32 points across the annotation task. Through six targeted analyses, we find that (i) label confusion concentrates on the negative/neutral boundary, (ii) two annotators show run-length drift consistent with autopilot labeling, and (iii) the dominant predictor of κ is temporal simultaneity: tweets labeled within one minute achieve κ= 0.98, while those labeled more than a day apart reach only κ= 0.65. Annotation speed and tweet-level linguistic features show no meaningful association with κ. We benchmark three open multilingual encoders and proprietary models (GPT-5 and Gemini) on three-class sentiment classification; fine-tuning yields gains of 29 to 43 macro-F1 points over pretrained baselines, with GPT-5 few-shot leading overall (62.2 macro-F1). We release the dataset, per-annotation timestamps, and analysis code to support reproducible quality auditing for future African language NLP resources.

Arabic large language models must refuse harmful prompts without over-refusing benign or sensitive prompts, yet a single refusal rate hides this trade-off. We evaluate it using benign refusal B and harmful-prompt refusal H, where H measures refusal rather than harmful compliance. Across five Arabic-capable models and 130 runs on the full human-written AraSafe set, refusal-only supervised fine-tuning (SFT) collapses toward blanket refusal, whereas selected mixed-SFT configurations reach H = 90% to 93% at B = 14% to 23%; four selected configurations exceed the H = 90% target in all three runs, while Fanar does so in two of three. Direct Preference Optimization (DPO) and inference guards change B and H differently across models rather than acting as uniform upgrades. In a blinded 300-response audit, annotator binary-refusal agreement is 89.0% (kappa = 0.78); Qwen3Guard and Aya Expanse 32B reach 88.7% and 91.0% accuracy, respectively, with no conclusive paired difference. Selected SFT raises H on Arabizi for all five models, but none reaches 90%, showing only partial transfer from Modern Standard Arabic. Overall, the results support model-specific operating-point selection: set a deployment target and retain only interventions that improve it.