全部动态
本页 40 条10月9日 周五
Workerville：迈向智能体安全的组织行为学解释
基于 LLM 的智能体如今会持续与环境交互，并受到用户指令、同伴消息和长期记忆等组织渠道的影响。现有安全研究已经考察了这些影响，但主要将其视为彼此独立的智能体组件。此类因素如何从统一视角共同塑造智能体的安全行为，仍未得到衡量。为弥合这一空白，我们倡导将组织行为学（OB）作为研究先进智能体安全性的框架，围绕智能体运行其中的关系结构，重新组织研究对象、理论基础和实验设计。我们首次将组织行为学中一个与安全密切相关的经典子领域——反生产工作行为（CWB）——系统形式化为 Agentic Counterproductive Behavior（ACB）。ACB 明确了三类组织前因变量（垂直的主管关系、水平的同伴规范和内部认知结构），并将其映射到三类反生产结果维度（未经授权的信息披露、破坏性操作和生产偏差）。为实现 ACB 的操作化，我们提出了 Workerville，这是一个受控基准，通过在共享任务上操纵组织条件，将 16 种组织配置应用于 210 项任务，生成 3,360 个挑战，并由经过人工验证的智能体评判器进行评估。通过对 6 个前沿 LLM 的基准测试，我们发现：（I）负面的组织前因变量组合后会产生非单调放大效应：在不存在负面前因变量时，未经授权的信息披露率为 16.5%；存在两个负面前因变量时升至 60.1%；存在三个时则回落至 50.3%；（II）智能体会再现人类 CWB 研究所预测的典型行为模式；（III）这些结果确立了 OB 作为智能体安全研究系统性框架的地位，并指向一项新的研究议程。
Sera：用于可靠且可解释电池健康预测的语义表示聚合
电池健康状态（SoH）预测对电池管理至关重要，但由于电池退化具有非线性以及不同电池之间存在异质性，仍面临挑战。现有数据驱动方法主要使用时序模型从数值型电池时间序列中学习，而更高层次的退化特征通常未得到显式表示。然而，这些特征能够为预测提供退化指导，从而支持可靠的预测，并使退化的影响更具可解释性。在本文中，我们提出 Sera，一种以退化语义补充时序建模的语义表示聚合框架。在电池领域知识的指导下，Sera 从时间序列中提取退化语义，并利用基于规则的知识和基于 LLM 的解释构建两种互补表示。这些表示经过独立编码，并通过门控聚合与时序模型学习到的表示进行融合。在主流基准数据集上，针对多个预测时间跨度和不同的时序模型开展的实验表明，Sera 能够持续提升预测性能，相较于时序基线模型，预测误差最多降低 37.3%，并增强泛化能力。反事实分析考察了预测结果如何响应退化语义的变化，以评估模型的可解释性。结果表明，在测试的各个预测时间跨度上，预测响应均与关键退化描述符的含义一致。综上，这些发现表明，结构化退化语义与有效的聚合方法能够提升预测准确性，并支持面向先进电池管理的可靠且可解释的电池健康预测。
Scaling to Tens of Thousands of Test-Time Iterations with Loop-Native Attention Residuals
In this paper, we argue that looped Transformers need their own residual connections to prevent performance degradation as the number of iterations grows. We observe that increasing loop iterations can reduce reasoning accuracy: noisy state updates overwrite correct intermediate deductions and even undo completed solutions. This leaves subsequent iterations to recover lost information from an already degraded representation: once an error arises in an earlier loop, often as a result of long-range propagation through the recurrence, later loops find it difficult to correct. In this paper, we introduce InfiLoop, a loop-native residual connection that learns which past computations to retain and how much to accept from each new update. InfiLoop combines content-based weighting with learned temporal decay to maintain a running summary of recurrent states. An exact streaming recurrence keeps its persistent aggregation memory constant as the loop count grows. The resulting adaptive update suppresses unreliable proposals and preserves useful intermediate states. Across extensive reasoning tasks, a 7M-parameter InfiLoop model outperforms existing recursive architectures, reaching 97.9% exact accuracy on Sudoku-Extreme, and 13.6% pass@2 on ARC-AGI-2. Notably, on Sudoku-Extreme, InfiLoop continues to improve with test-time looping beyond 20,000 effective steps, showing that added depth translates directly into stronger reasoning. Our code is available at https://github.com/pixeli99/InfiLoop.
记忆类型各异：通过多样化策略赋能 LLM 智能体实现长期记忆
近年来，大语言模型（LLM）的记忆能力日益受到关注。尽管现有方法已取得巨大成功，但基于检索的记忆方法通常忽视不同记忆之间的差异，采用统一策略处理所有记忆，从而导致性能欠佳。因此，一个直观的问题随之产生：能否将记忆划分为不同类型，并选择适当的策略？然而，鉴于记忆场景具有主题丰富、情境复杂和边界模糊等特点，要实现记忆的精准分类并不容易。为应对这一挑战，本文提出了一个记忆多分类数据集 TriMEM，为多样化场景中的记忆类型提供精准标注。在此基础上，我们提出了一种新颖的记忆框架 MemoType，该框架借助训练得到的路由模型，自适应地识别每条记忆和查询的类型。通过对记忆和查询进行路由，MemoType 能够检索与查询类型相匹配的记忆，并设计针对性的检索策略，从而提升检索性能。此外，我们从理论上证明，在多类别语料库中，任何单一检索策略的期望检索精度都存在一个基本上限，进而导致系统性的精度下降。在三个数据集上的大量实验表明，MemoType 始终优于现有方法，Recall@1 最高提升 16.18%。
SDPAD: A Fully Spike-Driven Pipeline for End-to-End Autonomous Driving
End-to-end autonomous driving demands trajectory planners that are both highly accurate and cheap enough for edge deployment. State-of-the-art artificial neural network (ANN) planners meet the accuracy requirement at the cost of heavy dense computation, while spiking neural networks (SNNs)---though promising orders-of-magnitude energy savings through sparse, event-driven arithmetic---still lag far behind in planning accuracy. We present \textbf{SDPAD}, a fully spike-driven end-to-end planning pipeline that closes this gap. SDPAD converts a pre-trained ANN perception stack into integer-spike form via quantized ANN2SNN conversion, lifts multi-view images into the bird's-eye-view (BEV) space with a spike-driven-max (SDM) depth distribution (Spike-3D-Lift), and plans through the Spike-QFormer, a spiking query transformer in which ego, agent, and map queries distilled from the BEV scene are fused by learnable waypoint queries via cross-attention, followed by deformable spike-cross-attention refinement. Every operation is gated by integer spikes and inference is a single feed-forward pass without temporal simulation loops. On the nuScenes open-loop benchmark, SDPAD achieves an average $L_2$ error of 0.40\,m and a collision rate of 0.12\%, on par with strong ANN planners while consuming 69.9\,mJ---less than 2\% of recent ANN baselines. In closed-loop evaluation on the NAVSIM navtest split, SDPAD reaches 86.3 PDMS, surpassing the previous SNN planner SAD by 4.3 points and matching mainstream ANN planners at a fraction of their energy. To our knowledge, SDPAD is the first fully spike-driven planner evaluated in end-to-end autonomous driving, demonstrating that SNNs can rival dense ANNs in complex driving tasks.
面向多语言 LLM 预训练数据选择的英语质量分类器适配
近年来，大语言模型（LLM）预训练的进展凸显了高质量训练数据在提升性能方面的作用。基于模型的过滤已被证明能够从 Web 规模语料库中有效筛选高质量子集，尤其适用于高资源语言；然而，低资源语言由于缺乏充足的标注数据而面临挑战。本研究探索将质量过滤扩展到 100 多种语言，提出了一种多语言适配方法，将现有的英语质量分类器转换为多语言版本。我们的方法以多语言文本为输入，以英语分类器应用于机器翻译文本后得到的分数作为标签，在 Transformer 仅编码器模型的嵌入之上训练一个小型多层感知机。我们的 1B、3B 和 8B 规模实验表明，该方法在不损害区域和文化知识基准表现的情况下，能够保持现有基于多语言模型的过滤基线在下游 LLM 基准测试中的性能。为进一步评估跨语言泛化能力，我们比较了高质量合成数据与 Web 样本的分类器分数，以及分类器分数与基于 LLM 的分数之间的相关性。结果表明，即使对于未包含在训练数据中的语言，该分类器也能学习其原始英语版本的评分标准。
Runnable Commit Untangling for Coding Agents
Coding agents produce large, tangled patches that mix multiple development purposes, making the code hard to review and maintain. Commit untangling offers the promise of organizing such large patches into untangled, manageable commits. This paper emphasizes two important limitations in existing commit untangling studies. First, they do not consider that untangled commits are ordered and should leave the code runnable. In practice, maintainers are unlikely to accept commits that prevent the code from running. Second, existing studies claim that commit untangling helps software maintenance. However, they conduct syntactic comparisons between the untangled commits and developers' original commits without directly showing the claimed maintenance benefits. To address these gaps, this paper makes two novel contributions: (1) RucTangle, the first agentic method that untangles commits while keeping the code runnable after each commit; and (2) TangleEval, the first evaluation framework that quantifies how untangled, manageable commit histories help coding agents repair bugs. We compare RucTangle against four untangling methods on 131 agent-generated patches. All histories produced by RucTangle are runnable, while baselines produce 20.6%-37.4% unrunnable commit histories. We further collect 453 agent-generated patches that introduce regressions (i.e., causing previously passing tests to fail) and ask two other coding agents to repair regressions. Augmenting agent context with RucTangle-produced histories yields 5.2% absolute improvement in pass@1. We also analyze agent trajectories to learn how they use untangled commits to navigate and fix bugs. Our findings demonstrate the value of adopting established software engineering practices in the era of coding agents, which broaden the future research agenda: how can agents actively use software history to make better development decisions?
大型语言模型的迭代削弱了对人工智能辅助科学写作的筛查
期刊和会议已开始对投稿论文进行筛查，以识别使用大型语言模型（LLM）生成的文本。这类筛查的可靠性依赖于针对一组固定 LLM 版本进行的基准评估，但实际使用的版本却在不断变化。本文量化了这种 LLM 更替如何影响科学论文的筛查。我们将《Proceedings of the National Academy of Sciences》中 4,000 篇 ChatGPT 出现前的摘要，与来自三家供应商的 23 个 LLM 版本所生成的改写文本配对；这些版本发布于 2023 年 6 月至 2026 年 8 月期间。随后，我们在不同的维护情景下训练检测器，从每出现一个新版本就重新训练的检测器，到训练一次后永不更新的检测器。在仅使用某一供应商过往版本训练的情况下，检测器可能在模型代际交界处性能崩溃：当其校准为误将 1% 的人工撰写摘要标记出来时，在最明显的交界点之前，能够检测出超过 99% 的改写文本，而在交界点之后仅能检测出 3.8%。使用较新版本训练的检测器也可能漏检较早版本生成的改写文本。不同版本之间的词汇差异，在很大程度上决定了检测能力可以迁移到何处，以及会在哪里失效。在我们模拟的两种筛查情景中，覆盖全部 23 个版本的筛查系统要么会将八分之一的人工撰写摘要标记出来，要么会漏掉最新版本改写文本中的三分之一。事实上，某款商业检测器漏掉了最明显交界点之后那个版本生成的大多数改写文本，却几乎没有将人工撰写的摘要标记出来。因此，科研诚信政策不应将检测器的基准准确率视为定论，而应将其视为暂定结果，并在每次 LLM 发布新版本时重新验证，包括对早期版本进行重新验证。
基于 LLM 的多智能体系统中的失败归因错误传播建模
基于 LLM 的多智能体系统（MASs）正日益用于通过协同推理、工具使用以及与外部资源交互来解决复杂任务。然而，对此类系统中的失败进行归因仍然具有挑战性，因为观察到的结果通常无法直接揭示导致执行失败的错误。在本文中，归因目标是决定性错误，即修正该智能体—步骤对便可恢复失败执行的错误。现有方法主要识别可疑步骤，却没有明确建模错误如何跨交互传播，或如何持续存在于尚未解决的循环中，因此难以将决定性错误与下游失败症状区分开来。我们提出错误传播建模失败归因方法（EMFA）。EMFA 构建失败轨迹的结构化表示，对级联传播和持续性交互循环进行建模，并采用具备传播感知能力的候选筛选，随后通过反事实验证来识别决定性智能体—步骤对。在 Who&When 基准测试中，EMFA 取得了当前最佳的步骤级归因准确率，并在智能体级别保持竞争力。与此前最佳的步骤级结果相比，它在 Hand-Crafted 和 Algorithm-Generated 子集上分别提升了 3.45 和 4.40 个百分点。
NanoProof：Lean 4 中开放且高效的自动定理证明
我们介绍了 NanoProof，据我们所知，这是 Lean 4 中首个因子化执行引导定理证明器，其训练数据、提取工具链、训练流程和权重均已发布，因此可以完全使用开源资源实现端到端复现。为此，我们构建并发布了一个结构化证明树数据集，以及一个用于在 Lean 4 形式化验证器中进行程序化交互和数据提取的工具。为支持可持续研究，我们注重计算效率，以促进易于获取的训练与评估。在 MiniF2F-Test 上，NanoProof 的 pass@16 达到 50.8%，超过了同类中最接近的两个系统 HyperTree Proof Search 和 ABEL，而所需计算量约分别少 90 倍和 7 倍；与 AlphaProof 相比，所需计算量则少四个数量级以上。虽然存在更强的开放权重证明器，但它们都是基于大型预训练语言模型进行微调的，既不发布训练数据，也不发布训练流程；NanoProof 表明，因子化执行引导类证明器可以使用适度资源从头重建。
Beyond Report Imitation: Clinically Aware Multi-Image Ultrasound Report Generation from Visible Evidence
Generating ultrasound reports from multiple images requires aggregating clinical evidence across views, yet archived key frames capture only part of the dynamic examination. Raw-report imitation is therefore misaligned with visual supervision: content that is clinically valid for the full examination may be unverifiable from the images available to a model. This gap creates a clinical behavior alignment problem. A model must preserve visible findings, avoid diagnostic reversals and unsupported completion, and not collapse into conservative templates. We propose CAMEO, a Clinically Aware Multi-image Evidence-grounded Orchestration framework for ultrasound report generation. Stage I learns ultrasound visual-language primitives; Stage II performs Cross-View Evidence Grounding by distilling trusted visible report points into multi-image QA and report-style supervision; and Stage III performs Clinically Aware Preference Alignment using clinical-error-oriented preference pairs. From USReport, we construct USReport-Distilled with 17,670 evidence-grounded paired-image training instances and USReport-Pref with 21,869 preference pairs; we additionally use 25,631 PubMedVision-US ultrasound instruction samples for domain adaptation and multi-image instruction tuning. On the primary USReport-Distilled benchmark, CAMEO improves over EchoVLM from 0.25 to 0.40 BLEU-1, 0.28 to 0.45 ROUGE-1, and 0.27 to 0.43 METEOR, while raising ClinicalScore from 55.02 to 74.20. These results underscore the value of evidence-grounded supervision, clinically aware alignment, and clinically structured evaluation for reliable ultrasound report generation.
AgentEvolver：通过任务执行实现系统级自我演化
智能体可以完成任务，却不一定会改进其工作方式。要将任务经验转化为可复用能力，需要把发生变化的组件与其评估及后续使用连接起来。我们提出 AgentEvolver，这是一种在任务执行过程中开发能力、同时保持基础模型不变的系统。八类实体通过一个统一的版本化生命周期，使可复用的操作、方法、智能体、控制流、接口和支持状态能够被修订。共享的 Runtime 协调持续进行的工作，而持久化规划和可恢复上下文则保留任务方向及支持性证据。我们在 SWE-bench Pro Public 上评估任务结果，并考察六个应用案例中的能力变化。研究团队报告称，引入演化后，任务解决率达到 82.08\%，超过其报告的未引入演化时的基线。案例显示，保留的能力被用于后续的网站、游戏和研究工作，同时也记录了未完成的目标和一次未成功的策略。这些发现区分了可复用组件的改进与最终任务的成功。AgentEvolver 为通过执行研究能力累积提供了具体基础；独立任务迁移和总开发成本仍是有待解决的问题。
TAM: Task-Aware Memory Distillation for Efficient Spatiotemporal Prediction
Knowledge distillation enables efficient spatiotemporal prediction by transferring knowledge from an accurate teacher to a compact student. However, matching outputs or features independently for each sample leaves cross-sample predictive structure underused. Exploiting this structure requires representations and historical references that reflect the dynamics of each task. We propose TAM, a Task-Aware Memory Distillation framework that organizes a frozen teacher's knowledge into a bounded, retrievable history. Memory entries encode latent features, forecast changes, or flow residuals, while task-specific selection rules identify relevant historical references. The student either matches the teacher's similarity distribution over shared references or regresses observation-conditioned residual prototypes. These objectives complement supervised prediction and conventional distillation. The teacher, memory, and auxiliary adapters are used only during training, leaving student inference unchanged. We evaluate TAM on video prediction, weather forecasting, and traffic flow prediction across multiple teacher-student configurations. Averaged over four paired runs, adding TAM improves SSIM on all six video datasets and reduces MSE on five relative to the corresponding KD baselines. Mean paired MSE reductions reach 1.86% on KittiCaltech, 1.93% on WeatherBench with a gSTA teacher, and 1.01% on TaxiBJ. These results demonstrate the utility of historical teacher supervision across distinct forecasting tasks without additional student inference cost.
适应位置至关重要：用于能力保留的层选择性微调
参数高效微调（PEFT）使大型语言模型（LLMs）能够适应专门任务，但往往会付出预训练过程中获得的通用能力退化的代价。现有方法主要通过数据回放或正则化来缓解这一权衡，依赖额外数据或显式的优化约束。我们转而关注另一个问题：应当在哪里进行适应？我们发现，对不同的 Transformer 层进行微调会带来不同程度的目标任务收益和能力退化，这表明并非所有层都同样适合进行适应。为刻画这种差异，我们使用逐层经验 Fisher 信息来衡量对目标任务的敏感性。然而，计算 Fisher 分数需要反向计算，并且对于大型模型而言成本会越来越高。因此，我们引入输入--输出余弦相似度，作为一种轻量级、仅需前向计算的层敏感性排序代理指标。在不同模型和任务上，输入--输出相似度较低的层始终表现出更高的经验 Fisher 分数。基于这一观察，我们提出 Layer-Selective LoRA（LS-LoRA），仅在输入--输出相似度较低的层中放置可训练的 LoRA 适配器。在数学推理和代码生成任务上的实验表明，与标准的全层 LoRA 相比，LS-LoRA 在保留更多常识推理能力的同时，能够提升目标任务的平均性能，说明谨慎选择适应位置可以成为平衡目标任务适应与通用能力保留的一种简单而有效的方法。
用于时间模式识别和机器人控制中动态手臂手势速度估计的神经网络
部署通过手势与人类互动的智能机器人系统，需要具备识别多样时间模式能力的神经网络。我们对十项抽象序列任务进行了系统性基准测试——其中五项为排列不变（集合）问题，五项为顺序相关（序列）问题——并在涵盖循环、卷积、基于注意力机制以及集合函数等类别的十八种神经网络架构上进行评估。除核心的架构—任务组合外，我们还探索了多种预处理方法和目标变量变换，共形成超过250种不同的实验配置。所有变体均在完全一致的条件下进行训练和测试（固定随机种子、共享超参数、共享数据划分），以确保比较的公平性和可复现性。在全部十项任务上的排名显示，有四种架构始终表现优异：BiGRU、TCN、Conv1D 和 GRUReLU；在基准测试设置下，它们的参数量均少于2,000个，足以支持实时部署。基于这一排名，我们将三种架构各具特色的顶尖模型（BiGRU、TCN 和 GRUReLU）应用于一个实际机器人学问题：根据骨骼关键点序列估计动态手臂手势的执行速度。我们在一个自建数据集上评估了三种速度定义方式（峰值计数、周期时间和平均尖峰间距）。该数据集包含八类与交通相关的手势，共有通过 OpenPose 记录的256,710帧数据。在峰值计数定义下，最佳配置取得了0.198的平均绝对误差，对应约5%的相对误差；周期时间定义下的相对误差约为4%，平均尖峰间距定义下的相对误差约为8%。这些结果表明，神经网络能够根据骨骼数据可靠地估计手势速度，为实现具备速度感知能力的手势控制机器人系统开辟了道路。
LTBD：可学习的信任边界分隔符，用于防御提示注入
大语言模型（LLMs）在复杂任务上表现出色，但仍极易受到提示注入攻击的影响：嵌入外部数据中的恶意指令可能覆盖用户意图。现有防御方法仍受限于需要对模型进行微调、易受自适应攻击影响，或依赖脆弱的手工设计提示。我们认为，这种脆弱性的一个根本原因在于缺乏对信任来源的明确表示。为此，我们提出可学习的信任边界分隔符（Learnable Trust-Boundary Delimiters，LTBD），这是一种轻量级防御方法，能够在保持LLM参数不变的同时，在输入中显式编码信任边界。LTBD使用少量可学习的分隔符来区分受信任的用户指令与不受信任的外部数据，使模型能够更好地遵循预期的信任层级。实验结果表明，LTBD显著优于推理时防御方法，并且与基于训练的方法相比具有竞争力，同时保持良性任务的实用性，仅引入可忽略的推理开销。具体而言，LTBD在AlpacaFarm上实现了0.00%的ASR，在TaskTracker上仅实现了0.11–0.19%的ASR。在自适应攻击下，LTBD同样保持有效；在这类攻击中，对手完全了解该防御机制，并明确尝试绕过它。
One Skill Too Many: How Co-Installed Skills Conflict in Coding Agents
Coding agents are extended with agent skills, directories whose SKILL.md tells the model when and how to perform a task. Because skills come from independent sources (teams, developers, plugins, copied collections), an installed skill can be co-installed with a similar skill doing the same job, and the model picks between them by name and description alone. In a conflict, the installed skill loses core functions (e.g., a ban on touching git) because the similar skill runs instead or changes what it does. The task still passes, so benchmarks that check only task completion miss such cases. We present the first empirical study of such conflicts. From snapshots of 20,947 repositories, we mine 822,109 candidate similar-skill pairs, have an LLM judge a stratified sample of 3,754, and run 312 confirmed pairs on three models (6,368 runs, 169,294 tool calls, 542 agent-hours). We report five findings. (1) Conflict-prone skills are common: nearly one in four installed skills is co-installed with one that does the same job, and 37% of judged skills sit inside copied collections. (2) Most such pairs involve normative skills, then capability skills. (3) Without lowering task completion, a similar skill takes one in five runs from the installed skill, and runs that open the similar skill first lose over a third of the exclusive core functions that only the installed skill fulfills. (4) Install location decides which skill runs, listing order barely matters, and the final reply names the skill used in only 0.9% of substituted runs. (5) Conflicts are decided at the first skill read, almost always before any file is changed, and a pre-tool hook at that read restores fidelity on exclusive core functions to the level of runs that open the installed skill first. Benchmarks should thus score exclusive core functions, and platforms should guard the first read and show which skill ran.
SkillContrast: Difference-Guided Text Selection for Agent Skill Reranking
Similar agent skills can share instructions but differ in their conditions of use. Query-based text selection may retain shared instructions and omit these distinctions. We introduce SkillContrast, a training-free selector that compares retrieved skills and retains their differing text with local context for a pretrained reranker. On 1,235 requests from SameCapRisk-Bench, it yields 54-72 more clean hits (requests that retrieve a helpful skill without its marked risky sibling) than TF-IDF query selection at identical per-candidate input lengths, across 2 retrievers and 2 reranker sizes. Length-matched component replacements identify differing text as the main contributor in the primary setting, with smaller, mixed context effects. Relative to full skill bodies, SkillContrast uses 51.1-58.8% fewer model-input tokens, with 10-18 fewer clean hits at 0.6B and matching or higher observed clean-hit counts at 4B. Candidate-relative differences thus complement query relevance in selecting compact reranking inputs.
Evidence-Traceable Dynamic Interviewer Architecture for Expertise-Adaptive Qualitative Interviews Using Local LLMs
Automated interviewers and conversational agents are increasingly used in research, recruitment, customer service, and education. However, many existing systems rely on fixed question sequences and provide limited context-based personalization without considering participants' knowledge, which can lead to repetitive or irrelevant follow-up questions. Therefore, there is a need for an adaptive interviewing system that can adjust question depth while maintaining conversational continuity and semantic progression. To address this, an Evidence-Traceable Dynamic Interviewer Architecture is presented using a locally hosted Large Language Model (LLM), with the interview continuously adapted throughout the entire conversation based on the participant's responses and evolving context. The interviewer profiles participants' expertise in real time to generate knowledge-appropriate questions, well-articulated responses, and smooth transition messages that support conversational continuity. A five-module prompt-driven architecture and persistent interview-state record support these functions. The interviewer was evaluated with 246 participants. Expertise Profiling module (M3) showed 78.9% exact agreement with independently reported participant expertise, with a weighted Cohen's K of 0.80. Generate Iterative Questions module (M4) showed a strong expertise-complexity association (p=.79, p<.001), and participants reported high relevance (mean 4.41), engagement (mean 4.32), and satisfaction (mean 4.38), providing evidence that the architecture's adaptive components operated consistently with their intended functions while participants reported a positive interview experience.
运行框架演化触及上限：何时应开始训练权重
要提升长时程 LLM agent，可以围绕冻结模型演化运行框架，也可以训练模型权重。我们先让自演化运行框架增强系统，再将初始和演化后的运行框架分别与基础权重和训练后权重组合，以了解训练后的模型保留了哪些增益，以及哪些增益仍然依赖运行时环境。我们发现，可以从 agent 的失败构成中判断应采取哪种手段：按照失败轨迹中最先触发的信号进行标注，可以将过程失败（调用受阻、循环、步骤预算耗尽）与内容失败（计划已交付但质量较差）区分开来。运行框架演化能够修复前者；它塑造的行为可以训练进权重；而内容失败正是权重训练要解决的问题。在 DeepPlanning 上，自演化运行框架循环将 Qwen3.5-4B 的留出集得分从 0.16 提升至 0.30，将 Qwen3.5-9B 的得分从 0.32 提升至 0.44；对于 4B，留出集交付率从 55% 提升至 90%，而内容失败则留待权重处理。在演化运行框架轨迹上训练的 LoRA 适配器能够将这一增益内化：在原始运行框架下，两种规模的模型在留出任务上都额外提升 +0.13；对于 4B，适配器与运行框架结合后，留出集得分提高到原来的两倍以上；对于 9B，仅使用适配器就能达到完整演化流程的效果，并将内容失败率从四分之一的轨迹降至二十分之一。一个在答案打乱后的轨迹上训练的安慰剂适配器，其表现低于基础模型。该循环还迁移到了 WebArena-Lite（在 117 个未见任务上提升 +0.09），其中增益来自模型所看到的内容，适配器无法进一步增加这一增益。由此得到一条两次应用的“先诊断、后干预”规则：先读取失败构成，在运行框架与权重之间做出选择；再观察被接受的编辑改变了什么，以决定应将哪些增益训练进模型。得分是相对于新锚点进行的四次 rollout 均值；除特别标注者外，评测均在同一晚完成，覆盖来自六个模型家族的八个模型和两个基准测试。
DIAL-OPD: Learning More from Fewer Tokens in On-Policy Distillation
On-policy distillation (OPD) supervises student-generated trajectories with token-level teacher signals. Its sampled-token variant avoids the cost of full-vocabulary probabilities. Yet we find that training on fewer tokens can outperform full-token OPD, challenging the intuition that more supervision improves learning. This motivates selecting tokens by learning value. Existing disagreement-based criteria ignore probability scale: tokens assigned negligible probability by both models, termed low-low tokens, can receive large log-ratio rewards and hinder learning. We propose DIAL-OPD, a token-selection method that bridges log-probability and probability spaces by weighting reward magnitude with the logarithmic mean of teacher and student probabilities. A parameter beta controls this weighting, and the highest-scoring tokens are retained. Across 4 teacher-student pairs and 7 mathematical reasoning benchmarks, we compare DIAL-OPD with 9 baselines. Retaining only 40% of tokens, it outperforms Vanilla OPD and its full-token variants, with mean accuracy gains reaching 5.25 percentage points over Vanilla OPD, and doubles AIME25 Pass@16 from 13.33% to 26.67%. It also achieves up to an 18% relative improvement in mean accuracy over the strongest token-selection baseline at matched retention ratios. With a 4B teacher, DIAL-OPD surpasses the strongest full-token baseline using an 8B teacher at both student scales, showing that effective supervision allocation can outweigh teacher scaling. Further analysis shows that moderate beta balances suppressing low-low tokens against preserving useful disagreements. Token-level evidence reveals that DIAL-OPD filters high-reward tokens with limited reasoning value while preserving supervision critical to reasoning correctness.
Lamarck的驾驶学校：通过进化竞争发现自动驾驶训练策略
自动驾驶能力在很大程度上取决于训练过程中所遇场景的分布。现有方法通常使用真实性、难度或风险等替代性标准来构建或动态调整训练场景分布。然而，这些预先设定的替代标准可能无法准确反映训练价值，导致训练资源使用效率低下。为解决这一局限性，我们提出了一种拉马克式进化框架，以候选分布之间的竞争取代基于替代标准的引导。我们将训练策略发现形式化为一个多阶段双层优化问题，并使用拉马克进化算法对其进行近似求解。在外层，达尔文式的交叉、变异与选择探索场景分布空间；在内层，策略学习获取新的能力，而拉马克式继承将这些能力传递至后续阶段，使场景分布与策略能力能够协同进化。由此产生的进化轨迹揭示了高价值分布之间反复出现的阶段性规律，其特征是通过阶段性挑战轮换实现能力积累。我们进一步将这些规律提炼为一种轻量且可复用的拉马克式训练策略。实验表明，与基线方法相比，完整框架最多可将性能损失降低25.07%，而轻量策略仍可实现19.13%的降幅。这些结果表明，进化竞争不仅能够发现有效的训练策略，还能揭示可复用的阶段性模式，即训练分布的价值如何随策略能力的变化而改变。代码已发布在GitHub上。
Constrained Command-Conditioned Reinforcement Learning with Bandit Strategy Selection in Real-Time Strategy Games
Deep reinforcement learning agents reach strong performance in real-time strategy games but can be brittle against opponents outside their training distribution. Separating strategic command selection from learned unit control allows different strategies to be selected for different opponents while reusing the same execution policy. This requires an executor that can follow different commands and measurable criteria for assessing whether it does so. We introduce a constrained command-conditioned Proximal Policy Optimization (PPO) policy, the executor, for MicroRTS, a real-time strategy environment. Discrete commands specify strategic objectives and behavioral requirements for economy, army composition, military posture, and worker policy over multiple environment steps; the executor determines the unit-level actions used to fulfill them. A Thompson-sampling bandit acts as the strategist, selecting command tuples from an estimate of the opponent's strategy built from in-game observations rather than opponent identity. In a controlled comparison with a flat PPO baseline trained with the same architecture, budget, curriculum and self-play league, the strategist-executor system wins significantly more often against three of the four strongest opponents on a training map, including the two strongest held-out ones (0.55 to 0.97 and 0.01 to 0.34), with no significant difference against the others.
HI3D 3.0 (Twinkle3D): Object-specific 3D Asset Generation with High Resolution
Image-to-3D generation has become increasingly capable of producing objects that closely resemble the input image, and an outstanding challenge is to reproduce the depicted object itself, including the specific geometry that defines it. Inscriptions, brand marks, and repeated structures are frequently distorted or lost, despite being critical to object identity. We present Hi3D 3.0, an image-to-3D generation system targeting object-specific fidelity, with Twinkle3D as its geometry model for generating watertight triangle meshes at $2048^{3}$ resolution. Twinkle3D advances high-fidelity geometry generation along four dimensions. First, while O-Voxel/FaithC offers high representational precision, it often suffers from poor surface quality and non-watertight geometry. We address both issues while retaining its $2048^{3}$-level precision. Second, we scale diffusion generation to sequences of up to 300K geometric tokens through a redesigned DiT architecture and large-scale distributed training optimizations, reducing training time per step from approximately ten minutes to ten seconds. Third, subsequent refinement cannot fully compensate for errors introduced during initial generation; we therefore strengthen both global shape and local detail in the initial generation stage, and the resulting single-stage model surpasses prior two-stage pipelines with $512^{3}$ refinement. Finally, we introduce a fine-grained image-3D cross-modal interaction mechanism that strengthens correspondence between visual evidence and geometric tokens, improving the recovery of object-specific structures. We evaluate geometric fidelity using alignment metrics derived from silhouettes and normal fields. Hi3D 3.0 outperforms four commercial systems across all reported metrics, recovering 82.1% of inscribed characters at 98.2% precision, compared with 21.7% recall for the strongest competitor.
Epistemic Disturbance in the Graph Model for Conflict Resolution: State-Preserving Actions, Four-Valued Assessments, and the Distinction between Capability and Intention
In the graph model for conflict resolution (GMCR), a decision maker (DM) either moves the conflict to another state or does nothing. The basic definitions leave inaction implicit, so every action that leaves the state unchanged is treated as doing nothing. Yet announcements, exercises, leaks and selective disclosures are neither moves nor inaction: they leave the state unchanged but change what other DMs believe about which moves are available and which moves others would want to make. We introduce such state-preserving actions by augmenting states with the DMs' epistemic states: a physical move changes the physical state, a state-preserving action changes only the epistemic state, and inaction is the absence of a transition. Actions generate evidence through observer-specific interpretation maps. Building on a four-valued extension of GMCR from the author's earlier work, which separates evidence for and against, we show that evidence for a move can only enable perceived moves and evidence against can only disable them, that two of the four reduction operators ignore one kind of evidence, and that contradictory assessments are absorbing under monotone accumulation. With the monotonicity of stability in move sets, this fixes the direction in which any action moves a DM's stability judgements and characterizes when actions can enable provocation or deterrence. Capability assessments affect all sanction-based stability concepts, and on the DM's own side also Nash stability, whereas intention assessments affect only sequential stability. Hedging between two candidate types weakly expands or shrinks an observer's sequentially stable set according to how it reads contradiction. In the 1995 DVD format negotiation, general metarationality cannot distinguish its phases, since the computer industry group could always sanction; sequential stability, which asks whether it would, can.
Can Jev be Your Q or Policy in Reinforcement Learning?
Foundation models supply reinforcement learning (RL) with priors that mitigate its longstanding weaknesses in sample efficiency and transfer, but their token-by-token generation makes queries sequential and costly. Jev, a recently released decision model, generates nothing and returns calibrated, typed answers in a single forward pass. Existing work studies foundation models in RL either as models to be trained or as generators to be prompted, and Jev belongs to neither category, having so far served only as a black box in single domains. How well such a model decides on its own in RL environments, and how it can improve RL as a component of training, therefore remain unaddressed. To this end, in this paper we first examine the requirements that the objects of an RL system place on the answers they consume, and establish that Jev can fulfill all of them except the cardinal use of a value function. The remaining objects form positions that admit several roles each. We then construct algorithms that employ Jev at three of these positions, as a reference policy, an exploration judge, and a replay rater, to improve sample efficiency, exploration, and learning performance. Across nine MiniGrid tasks and three Atari games, training with Jev outperforms a standard RL learner, including where the learner makes no progress alone, while the model itself remains untrained. To our knowledge, we present the first use of Jev within the RL learning process and establish a frozen decision model as a usable component of RL training, inviting further exploration of how Jev and other advanced decision models can improve RL.
Elucidating the Space of Enzymatic Reaction: A Unified Benchmark and Pretrained Model
Existing reaction models primarily learn molecular transformations, whereas enzy- matic reactions depend jointly on molecular structure and catalytic function. We formulate this problem as learning an enzymatic reaction space linking reactants, products, and Enzyme Commission (EC) annotations. To characterize this space, we introduce VenusRX-Bench, a unified benchmark for forward reaction prediction, single-step retrosynthesis, and EC-number prediction. VenusRX-Bench integrates reactions from multiple biochemical databases with standardized curation, leakage- controlled splits, and consistent evaluation. Benchmarking representative chemical and enzymatic models reveals a clear chemical-to-enzymatic domain gap, driven by limited domain data, catalytic-context dependency, and the difficulty of modeling large biomolecular structures. To bridge this gap, we develop VenusRX, a unified T5-style sequence-to-sequence model for enzymatic reactions. VenusRX jointly learns forward prediction, ret- rosynthesis, and reaction reconstruction, with two-stage training on millions of template-expanded reactions followed by real biochemical reactions. In addition, optional EC conditioning incorporates catalytic context, while Molecule Library- Constrained Decoding improves the generation of complex biomolecules. Across benchmark tasks and challenging generalization splits, VenusRX achieves the best or competitive performance on most evaluated settings over representative chem- ical and enzymatic baselines. Moreover, EC information consistently improves reaction prediction, while learned reaction representations support accurate EC prediction, revealing a bidirectional relationship between reaction structure and catalytic function. Together, VenusRX-Bench and VenusRX provide a unified framework for elucidating and modeling enzymatic reaction space
A 3D Characterization Framework for Intelligent Sequential Decision Making
Puzzles are widely used to evaluate the reasoning capabilities of artificial intelligence (AI) systems for sequential decision making, yet approaches originating from different paradigms are rarely compared under unified conditions. To address this gap, we introduce a three-dimensional characterization framework that enables the analysts of AI methods by 1) projecting them to the Markov decision process (MDP) sequential decision making formalism, 2) degree of autonomy through human prior ranking of their designs and, 3) skill and computational cost. Using this framework, we analyze how representative graph-based, reinforcement learning, and large language model (LLM)-based approaches differ in their design choices and performance characteristics, instantiated respectively by Neurosolver, forward-backward reinforcement learning (FBRL), and automated thought-of-search (AutoToS), including a double-agent extension of thought-of-search (DA-ToS). The analysis relies on the Tower of Hanoi puzzle that provides a controlled benchmark with well-defined rules and scalable complexity, enabling consistent comparison across increasing problem sizes. The 3D characterization reveals that LLM-based methods, due to their weakly constrained action-space design, shift complexity from architecture to inference-time verification, leading to substantially higher memory and runtime costs than Neurosolver and FBRL.
Intervention anchors and scientific verification in synthetic vascular predictive representations
Complete orthogonal predictive coordinates do not by themselves bind a latent direction to a named intervention. We present a mathematical and synthetic audit motivated by vascular device-vessel suitcordance. Capacity-matched least-squares predictors were exactly equivalent under complete fixed output transforms, whereas an anchor-only observer recovered interpretations only within the span of known perturbation signatures. Six three-dimensional configurations across 64 seeds gave a maximum paired prediction discrepancy of 6.7e-15 but a median untransported edit error of 1.513. Coordinate transport removed that error. Noisy and weak anchors constrained calibration stability, and changing the representation basis required recalibration or verified transport. Across 256 additional fits in dimensions 3-24, prediction equivalence persisted within 4.0e-15. We then evaluated nine deliberate runnable fault classes across 64 seeds. All 576 faulty executions completed, but each violated at least one reconstruction, prediction, delivered-edit or scope contract; all 320 valid control records passed. Repeating a faulty implementation gave exact self-agreement despite error against the separately computed simulator expectation. For one omitted-direction defect, probe coverage followed its analytic law, and rank-aware abstention protected unsupported interpretations. Scalar-noise experiments exposed both missed weak faults and excessive rejection under narrow relative tolerances. These controls provide an executable separation of prediction, semantic support and scientific acceptance. They are synthetic numerical audits, not clinical validation, neural JEPA-Anything replication, agent learning or patient treatment-effect estimation.
DeltaReplay: Task-Relative Memory Reuse for Mobile GUI Agents
Memory-augmented mobile GUI agents store successful execution trajectories and reuse them in later tasks, but a stored trajectory rarely matches a new task exactly. The new task may use different parameters, share only some of its steps with a stored trajectory, or have no relevant record in memory. Forcing the agent to use irrelevant memory can mislead it, whereas discarding memory that may still be useful deprives it of guidance from past experience. To address this dilemma, we propose DeltaReplay, a step-level memory reuse framework that decides how to use existing memory without modifying it. We observe that the reusable part of a stored record is determined not by the record itself but by its relation to the new task, mainly through two factors: page-level consistency and action-level generality. We therefore store execution trajectories as paths in a transition graph, whose nodes (pages) and edges (actions between pages) capture these two factors. At reuse time, the action on each edge is split into a task-independent operation and task-specific parameters. DeltaReplay then compares each recorded step with the new task and the current screen, and decides whether to follow it, execute it after replacing its parameters, or leave it to the base agent. On AndroidWorld and SPA-Bench, DeltaReplay improves the task success rate over a base agent with the same backbone by up to 10.3 and 25.0 percentage points, respectively. These results indicate that deciding at each step how to use retrieved memory lets agents benefit even from partially matching trajectories.
CoreWeave targets AI inference bottlenecks with full-stack optimization
AI inference is fast becoming the workload that decides the economics of the AI boom. Training built the first wave of GPU clouds, but serving models faster and cheaper will define the next. That shift is pushing specialized cloud providers beyond raw GPU capacity into storage, networking and software. One provider is layering managed services […] The post CoreWeave targets AI inference bottlenecks with full-stack optimization appeared first on SiliconANGLE .
ttok 0.4
Release: ttok 0.4 ttok is my CLI tool for counting tokens, using OpenAI's open source tiktoken library. It hasn't been in updated in a couple of years, but I finally fixed a Click warning, updated CI, and added a --list-models command to list available models. It works with uvx , so you can count tokens in anything like this: cat file.txt | uvx ttok Tags: projects , ai , openai , generative-ai , llms , tokenization