2026-09-08

Bilevel Coordinated Reflection: A Game-Theoretic Approach to Multi-Agent LLM Systems

👍 130

09/01 20:00

Multi-agent LLM systems commonly use an orchestrator to decompose a task for a team of workers and then improve through textual reflection. Despite strong empirical results, these systems lack a unified account of coordination, memory improvement, and the role of external verification. We model orch

中文介绍 论文提出了一种双层协调反射的方法，旨在通过博弈论框架提升多智能体 LLM 系统的协调性和内存改进。该方法建模了任务分解和文本反思过程，强调外部验证在提升系统表现中的重要作用。此研究为多智能体协作和记忆增强提供了统一的理论支持，推动了在复杂任务处理中的应用。意义：对 Agent 和协作系统的研究方向具有重要影响。

Dr. Claw: An AI Scientist Workspace for Vibe Research

👍 125

08/30 20:00

Command-line coding agents (e.g., Claude Code, Gemini CLI) can already read and write files and sustain long sessions, yet end-to-end research still fragments across chat tools, IDEs, terminals, and writing environments, and the decisions that make it auditable are rarely preserved. We present Dr. C

中文介绍 Dr. Claw 提供了一个用于 Vibe 研究的 AI 科学家工作空间，旨在整合目前分散在聊天工具、IDE 和终端中的研究过程。该系统借助命令行编码代理提高研究效率，解决了信息碎片化和可审计决策缺失的问题，为科研人员提供了更加连贯的工具支持。意义：有助于推动科学研究中 Agent 和交互工具的整合与发展。

Unlocking Lossless Speedups in LLMs via Discrete Diffusion

👍 112

09/02 20:00

Large Language Models (LLMs) owe much of their success to next-token prediction (NTP), but their autoregressive (AR) structure requires slow, sequential token generation. To overcome this bottleneck, we introduce diffusion-augmented LLMs, a new class of models that defines an AR model distribution w

中文介绍 本文介绍了一种新的模型类，称为扩散增强 LLM，通过离散扩散解决了大型语言模型在自回归结构下生成速度慢的问题。该方法允许更快速的序列生成，同时保持模型的预测性能，指出在生成速度和精确度之间找到平衡的重要性。意义：推动了 LLM 的应用效率，可能影响实时生成任务的发展方向。

FlowBalance: Verifier-Grounded Self-Improvement from On-Policy Reasoning Experience

👍 81

09/02 20:00

A reasoning model can improve from its own on-policy experience, but this inner loop is fragile: terminal verifiers provide reliable yet sparse supervision, while dense same-model guidance can reinforce false confidence or overconcentrate learning on a narrow solution mode. We introduce FlowBalance,

中文介绍 FlowBalance 通过引入终端验证器，在机遇与风险之间为推理模型的自我改善提供了一种新方法。该方法集中于利用差异化的反馈机制以实现更有效的学习，解决了密集模型引导可能带来的错误信心问题。意义：为推理和自我改进策略提供了新的视角，对算法健壮性提升具有重要意义。

Iris: Climbing to the Search Frontier

👍 54

09/02 20:00

We present Iris-mini and Iris-pro, two search agents trained at the 35B-A3B and 397B-A17B scales, together with the data pipeline and training recipe behind them. Tasks are reverse-constructed from the hyperlink structure of a web corpus: we author multi-hop chains over an entity graph distilled fro

中文介绍 论文提出了 Iris-mini 和 Iris-pro 两种搜索代理，分别在 35B-A3B 和 397B-A17B 规模上进行训练，采用从网络语料的超链接结构逆向构建任务的方法。此研究展示了在多跳链条上的有效搜索能力，为信息检索体系的设计提供新思路。意义：推动了搜索算法的创新与应用，尤其是在复杂数据环境中的搜索效率提升。

Scal3R: Learning Efficient Multi-Relative Pose Query for Scalable Online 3D Reconstruction

👍 46

09/02 20:00

Online 3D reconstruction models perform poorly on long videos. This happens because regressing poses relative to a fixed first-frame anchor forces extrapolation far beyond the training distribution. Small drifts accumulate and amplify into significant geometric collapse. However, we observe that per

中文介绍 Scal3R 通过学习高效的多相对位姿查询，实现了可扩展在线 3D 重建，解决了在长视频中因固定第一帧锚点而导致的位姿回归问题。文章指出通过相对位姿的引入有效减少了几何崩溃现象，提高了长期视频处理的精度。意义：为在线 3D 重建和计算机视觉领域提供了新的解决路径。

Motion-Omni: End-to-End Joint Speech and Full-Body Motion for Spoken Dialogue

👍 38

08/27 20:00

An avatar that holds a conversation should decide what to say and to move while saying it, yet these abilities live in separate model families: spoken dialogue models produce speech without motion, and co-speech motion models produce motion only from audio handed to them. The standard remedy is a ca

中文介绍 Motion-Omni 提出了一个端到端的联合语音与全身动作生成模型，解决了传统对话模型与动作生成模型分离的问题。该模型实现了语音与动作的同步生成，提升了虚拟形象在对话中的表现力。意义：对人机交互和虚拟角色技术的发展具有重要的促进作用。

ENEAS: Embedding-guided Neural Ensemble for Adaptive Segmentation

👍 35

09/02 20:00

We present ENEAS, a unified, text-promptable method for instance tracking and semantic discovery. Text-promptable segmentation models, including the latest foundation models such as SAM 3, still suffer from temporal hallucinations, spatial fragmentation, and semantic misclassification: they fail to

中文介绍 ENEAS 提出了一个统一的文本提示式实例追踪与语义发现方法，旨在解决现有模型在时间和空间上的困难，如时间幻觉和语义错误分类。该方法通过增强文本提示的设计，提升了实例分割的精确性与有效性。意义：对实例追踪和语义理解的研究提供了新的视角，促进了计算机视觉领域的进步。

The Attention Triangle in Audio-Video Models

👍 30

09/02 20:00

Audio-video diffusion models rely on cross-modal attention to coordinate text, sound, and visual content, yet this same mechanism can introduce subtle and systematic semantic leakage. We study these models by probing and analyzing the ``attention triangle,'' comprising the three cross-attention edge

中文介绍 本研究分析音频-视频扩散模型中的交叉注意力机制，强调其在协调文本、声音和视觉内容时可能的语义泄漏问题。通过对“注意力三角形”的深入探讨，揭示了不同模态间信息处理的复杂性与潜在风险。意义：为多模态模型的设计与优化提供了重要参考，有助于提高模型的安全性与可靠性。

DRACO: Fine-Grained Credit Assignment with Dynamic Rubrics for Long-Horizon Agent Training

👍 26

09/02 20:00

Reinforcement Learning from Verifiable Rewards works well when a task has a programmatic checker, but most long-horizon agent domains have none. We work in the outcome-blind setting, where ground-truth success signals are not available. Multi-criteria rubrics are a popular way to supply such a rewar

中文介绍 DRACO 研究了一种动态评分系统，以增强长时域代理训练的细粒度信用分配。在缺乏明确给定成功信号的环境下，该方法引入了多标准评分法，优化了奖励信号供给，推动了强化学习在实际应用中的可行性。意义：为长时域智能体的训练与评估提供了新的策略，促进强化学习的长远发展。

Enoki: Efficient Multi-Level Hallucination Detection

👍 23

08/31 22:21

Ensuring factuality remains a critical challenge for deploying LLMs in high-stakes settings. Existing hallucination detectors usually operate at a single level: claim-level methods provide interpretable factual units, while span-level methods localize unsupported text. Bridging these views is costly

中文介绍 Enoki 提出了高效的多级幻觉检测方法，致力于保障大型语言模型的事实准确性。通过结合声称级和跨度级幻觉检测方法，该系统能有效缩小现有检测方法的局限性，提高LLM在高风险场景下的实用性。意义：为语言模型的安全部署提供可靠的检测工具，增强其在关键领域的应用能力。

Using Grounded Theory for Agent Behavior Analysis at Scale

👍 19

08/30 20:00

Understanding agent behavior requires methods that scale to thousands of trajectories and surface new patterns in long, often unfamiliar tasks where pre-built classifiers fall short. We propose to bring grounded theory into agent trajectory analysis: a six-decade-old qualitative method from the soci

中文介绍 研究提出了一种将基础理论应用于代理行为分析的新方法，旨在应对数以千计的轨迹分析与长任务模式识别的挑战。该方法通过结合定性方法与定量分析，改善了代理轨迹分析的可扩展性和有效性。意义：为代理研究提供了一种全面的分析工具，推动了智能系统理解的深入发展。

Ask Before You Optimize: Dynamic Pre-Formulation Clarification for Interactive Optimization

👍 18

09/03 20:00

Large language models (LLMs) are increasingly used to formulate optimization models from natural-language problem descriptions, yet realistic operations research (OR) requests are often incomplete: missing objectives, constraints, or business rules can change the resulting mathematical program. Exis

中文介绍 论文提出了一种动态预先公式化澄清的方法，以解决大型语言模型在处理优化模型时常遇到的信息不完整问题。通过引入实时澄清机制，该方法保证了最终数学模型的准确性，提高了优化决策的有效性。意义：促进了优化技术和语言模型结合的深入发展，推动了智能决策支持系统的应用。

Don't Drop Dropout: Optimizing Layer Sparsity for Efficient LLM Training and Inference

👍 18

09/03 20:00

Layer dropout (a.k.a. stochastic depth) has been shown to enable faster training, higher accuracy, and robustness to zero-shot layer pruning in both language and vision transformers. However, as models and datasets have scaled, dropout - particularly layer dropout - has largely disappeared from larg

中文介绍 研究探讨了层稀疏性优化对高效 LLM 训练和推理的影响，指出在大模型和数据集扩展情况下，传统丢弃法（dropout）在提高训练速度和准确性方面的重要性。此工作提供了对高效训练策略的新见解，倡导重新关注原有技术。意义：针对 LLM 的训练效率改进，推动了深度学习领域对模型优化的探索。

Select, Compress, Reinvest: A Controlled Study of Visual-Token Allocation in Long-Video MLLMs

👍 15

09/02 20:00

Long-video language models cannot look at every frame: an hour sampled once per second is 3,600 images, and a system keeps only a small fixed slice of that pool. Which frames survive that slice is usually treated as a preprocessing detail; we test whether it should be. Published selectors make the c

中文介绍 本文探讨了在长视频 MLLMs 中视觉标记分配的有效性，通过控制采样框架以提升信息处理效率。研究发现，优化选择机制对于提高模型表现至关重要，对视觉内容的选择与处理方式进行了系统化的验证。意义：为长视频处理和多模态模型的设计提供了新思路，推动了计算机视觉与自然语言理解的结合。

Beneath the Surface of Chains-of-Thought: A Mechanistic Interpretation of Reasoning Operations in LLMs

👍 14

09/03 20:00

Reasoning in large language models unfolds through diverse functional operations, such as problem formulation, goal decomposition, and deduction. Although these operations are explicitly distinguished in text, little is known about how they are geometrically organized in representation spaces. To th

中文介绍 文章通过机制解释法探讨了 LLM 中推理操作的多样性，研究了问题构建、目标细分和推理等操作在表示空间中的几何组织方式。该研究揭示了推理过程中的潜在结构，为进一步理解 LLM 的工作原理提供了基础。意义：对推理和思维模型的设计具有重要的指导作用，促进了机器推理能力的研究。

QCell: Recombining and Aligning Cell Queries for Overlapping Instance Segmentation

👍 14

08/28 20:00

Instance segmentation of overlapping cells in microscopy remains challenging due to semi-transparent structures that produce weak boundaries and mixed visual evidence in overlap regions. Existing methods address this through local regions of interest or shape priors but lack global reasoning across

中文介绍 QCell 研究了显微镜中重叠细胞实例分割的挑战，提出了一种重新组合和对齐细胞查询的方法。该方法旨在克服在重叠区域中检测语义边界模糊的问题，强化了局部与全局推理的结合。意义：推动生物医学图像分析技术的发展，提高了细胞分割的精确度与可靠性。

MaxKernel: Agentic Kernel Generation for TPUs

👍 12

09/02 20:00

Designing and authoring high-performance custom kernels for accelerators is a complex task that requires deep hardware-level expertise. Large Language Models (LLM) can be leveraged together with real-time compiler feedback to build agentic systems for kernel generation. In this work, we present MaxK

中文介绍 MaxKernel 探讨了基于 LLM 的高性能自定义内核生成方法，结合实时编译反馈，简化了为加速器设计内核的技术挑战。该研究展示了如何利用 LLM 在硬件层级开发中增强创造性支持。意义：对高性能计算和深度学习框架的优化具有重要的应用前景。

VeriPhy: Agentic Physical Reasoning for World Model Evaluation and Refinement

👍 12

09/01 20:00

Visual fluency in generated video does not imply physical reliability, and a scalar quality score alone is incapable of indicating the obligation a clip violates or the moment it fails. We present VeriPhy, an auditable physical-verification system in which a text-only planner compiles the prompt int

中文介绍 VeriPhy 提出了一种审计性的物理验证系统，通过文本规划器将提示编译为物理推理模型，专注于生成视频的视觉流畅性与物理可靠性之间的关联。该方法确保了生成内容的可靠检查，为生成模型的应用提供了验证基础。意义：提升了生成模型在逻辑推理方面的准确性，推动计算机视觉与科学推理的结合发展。

Causal Foundation Models

👍 11

09/01 20:00

Causal inference is the practice of estimating the effect of a treatment or intervention from data. It traditionally requires a bespoke pipeline for every new problem: first proposing a causal mechanism, selecting a compatible estimator, and finally training it. Meanwhile, across diverse settings an

中文介绍 Causal Foundation Models 研究了因果推断在多场景下的应用，提出了一种通用管道以处理每个新问题。该方法结合了因果机制的提议、合适估计器的选择及其训练，将复杂的因果推断过程变得简化和可重复。意义：为因果推断研究提供了统一框架，促进其在不同研究领域中的广泛应用。

GPT-6 Astra should be Illegal for SEO

@borjafat · 4.9K 粉丝 · 373.2K 阅 · 502 赞 · 42 转

09/07 07:52

If you want more customers from Google and AI search, but: ↳ ChatGPT keeps giving you keyword lists while your existing pages still need work. ↳ You open Search Console, stare at the numbers, and

中文介绍 博主讨论了使用 GPT-6 Astra 进行 SEO 的潜在问题，指出 ChatGPT 提供的关键词列表无法解决现有页面的问题，强调了 Google Search Console 中数据分析的重要性。这反映出许多人在 SEO 过程中依赖 AI 工具的现状和挑战。

Claude Code → Codex (and keep everything you already built)

@charliejhills · 14.1K 粉丝 · 120.8K 阅 · 504 赞 · 53 转

09/06 06:44

For months people have been telling me Codex is better than Claude Code. I ignored all of it. I have built my whole business inside Claude Code (every pipeline, every system) and I did not fancy

中文介绍 博主分享了将 Claude Code 转换为 Codex 的过程，并强调自己在 Claude Code 上建立的完整业务架构。他对改变平台的犹豫反映出对工具选择的深度思考，而这一点在行业讨论中越来越普遍。

Reducing cost and improving performance with Claude Platform

@ClaudeDevs · 683.4K 粉丝 · 106.4K 阅 · 1.3K 赞 · 70 转

09/08 13:01

Tuning prompt caching, instructions, and effort can reduce Claude's cost without sacrificing application performance. Performance and cost are often viewed as a trade-off: to spend less, you accept

中文介绍 博主探讨了如何通过调优提示缓存和指令来降低 Claude 的使用成本，同时提高应用性能。内容指出性能与成本并不是简单的权衡，反映出行业对优化 AI 工具使用效率的关注。

How to Use GPT-6 Astra to Find Profitable Trading Strategies 24/7

@RohOnChain · 68.8K 粉丝 · 102.3K 阅 · 524 赞 · 48 转

09/08 08:39

I will break down exactly how to build an AI agent like hedge funds that finds profitable trading strategies from the markets 24/7 with GPT-6 Astra. Let's get straight to it. Bookmark This - I'm

中文介绍 博主详细讲解了如何使用 GPT-6 Astra 构建一个 24/7 的 AI 交易代理，以寻找盈利的交易策略。这种方法与金融行业中 AI 技术应用的增长趋势相呼应，表明对 AI 交易工具的关注持续增加。

Using Grok Bot: Figma Photomosaics

@mattyp · 60.1K 粉丝 · 82.8K 阅 · 515 赞 · 56 转

09/07 18:03

You may have seen some three-panel mosaics on your timeline. Here's how you can create them with @bot in @figma. Any designer knows how they'd make one, it's just a bit of a pain in the ass. So we

中文介绍 博主指导如何在 Figma 中使用 Grok Bot 制作三-panel 的马赛克图。他指出这是设计师常用但繁琐的过程，这种自动化工具的分享反映出设计领域对简化工作流的普遍需求。

Exactly how I edit videos with Astra

@ForwardEditor · 10.0K 粉丝 · 61.4K 阅 · 512 赞 · 36 转

09/07 20:54

This weekend, I watched a movie with my daughter while my computer worked on a video I would have normally spent most of a day editing. The video was a favor for someone in my personal life. I'm a

中文介绍 博主介绍了如何使用 Astra 自动化视频编辑，大幅节省时间。内容分享了个人经验，体现了自动化技术在内容创作中的应用和提升效率的潜力。

The Math Needed for AI Inference Engineering (Complete Roadmap)

@TheVixhal · 23.0K 粉丝 · 59.7K 阅 · 541 赞 · 58 转

09/07 13:07

In this article, I'm going to break down the essential math you need for AI inference engineering. I'll also share the exact roadmap and resources that helped me personally. Let's get straight to it.

中文介绍 博主阐述了进行 AI 推理工程所需的关键数学知识，并提供完整的学习路线图和资源。这一内容对正在学习 AI 工程的从业者有重要参考价值，也反映出对基础知识体系的重视。

Grok Bot Marketplace is live!

@ericzakariasson · 87.8K 粉丝 · 58.2K 阅 · 522 赞 · 49 转

09/07 07:05

We launched the Grok Bot Marketplace last week on Grok Bot · Bot Marketplace! It's a shelf of teammates other people already built for real jobs. Pick one, import it into your sidebar, and you're

中文介绍 博主宣布 Grok Bot 市场的上线，提供了用户可以导入的已有团队机器人。这一举措代表了对低门槛 AI 工具和团队合作解决方案的日益关注，市场需求正在增长。

AlphaGenome Atlas: A High-Resolution Map of Human DNA Accelerating Biological Discovery

@GoogleAI · 2.5M 粉丝 · 37.4K 阅 · 643 赞 · 90 转

09/08 10:41

Today, we’re introducing AlphaGenome Atlas 🧬 an interactive platform for genomics discovery. In 2025, @GoogleDeepMind released AlphaGenome to help decode the complex language of our DNA — but

中文介绍 博主介绍了 AlphaGenome Atlas，一个加速生物发现的高分辨率人类 DNA 地图。该平台的发布标志着基因组学探索领域的重要进展，反映出对生物技术与 AI 结合的强烈关注。

How GPT-5.6 Sol helps run quantum computing experiments

OpenAI News · 09/08 13:00

See how an MIT researcher uses GPT-5.6 Sol with Codex to autonomously run quantum computing experiments, analyze results, and calibrate qubits.

中文介绍 麻省理工学院的研究者利用GPT-5.6 Sol和Codex自主运行量子计算实验，分析结果并校准量子比特。

Safety for Whom? Refusing the Right Subset of a Topic, Not the Whole Topic

Hugging Face Blog · 09/08 10:23

中文介绍 该条目暂无具体内容。

AlphaGenome Atlas: A predictive map of every possible DNA letter change in the human genome

DeepMind Blog · 09/08 10:00

AlphaGenome Atlas maps the molecular effects of 9 billion single-letter DNA variants across the human genome.

中文介绍 AlphaGenome Atlas映射了人类基因组中90亿个单个DNA变体的分子效应。

The Work Now Within Reach

OpenAI News · 09/08 09:00

Explore how more capable, affordable AI can expand the work people and businesses can accomplish—and make growth more economical.

中文介绍 探讨更具能力和经济性的人工智能如何扩展个人和企业的工作能力，并使经济增长更为可持续。

Introducing ChatGPT Images 2.5

OpenAI News · 09/08 07:30

ChatGPT Images 2.5 helps turn your ideas, sketches, and reference photos into more personalized, polished images that better reflect your ideas.

中文介绍 ChatGPT Images 2.5帮助用户将想法、草图和参考照片转化为更个性化、精致的图像。

This AI entrepreneur is developing agents that can plan ahead for the unexpected

MIT Tech Review AI · 09/08 06:34

Danijar Hafner’s office in San Francisco’s SoMa district sits mostly empty. His brand-new startup is still in stealth mode and doesn’t even have its name on the door. On the day I visit, there’s only one other person there, and little in the way of furniture. But what it lacks in decor, it makes up…

中文介绍 位于旧金山SoMa地区的AI企业家Danijar Hafner正在开发能够预判意外情况的智能代理，目前其新创企业还处于保密阶段。

On the Navier–Stokes Millennium Prize Problem

OpenAI News · 09/08 06:00

We’re sharing an AI-generated solution to the Navier–Stokes Millennium Prize Problem, including a writeup and a formal proof in Lean.

中文介绍 我们分享了针对Navier–Stokes千禧年大奖问题的AI生成解决方案，包括论文和在Lean中的正式证明。

Funding grants for new research into AI and teen development

OpenAI News · 09/08 05:00

Apply now for OpenAI’s $5 million grant program supporting independent research on how generative AI affects teen development, well-being, and safety.

中文介绍 OpenAI现开放500万美元的资助项目，支持独立研究生成性AI对青少年发展、安全与福祉的影响。

OpenAI expands initiatives to support journalism from classrooms to newsrooms

OpenAI News · 09/07 20:00

OpenAI is expanding support for journalism with tools, training, and partnerships for students, educators, journalists, and news organizations.

中文介绍 OpenAI扩大对新闻业的支持，包括为学生、教育者和新闻机构提供工具、培训和合作关系。

1Password increases engineering productivity 21% with Codex

OpenAI News · 09/07 20:00

Engineers at 1Password use Codex to rapidly build new features and internal tools, reaching production-readiness while maintaining rigorous security policies.

中文介绍 1Password的工程师利用Codex快速构建新功能和内部工具，提高工程生产力21%，并严守安全政策。

OpenAI managed agents 🤖, TPU inference ⚡, Anthropic $517B compute 💰

TLDR AI · 09/07 20:00

中文介绍 该条目暂无具体内容。

The Frontier AEO Tracker: What Astra Chooses (and every other frontier model, and what you can do about it)

Latent Space · 09/07 17:32

Our first Astra project dives into AEO trends, a top asked topic from founders and DX leaders we talk to.

中文介绍 我们的首个Astra项目探讨了创始人和DX领导人关注的AEO趋势。

Supporting independent journalism in Ukraine

OpenAI News · 09/06 20:00

OpenAI, AIRPPU and WAN-IFRA launch an AI program to help Ukrainian news organizations strengthen innovation, resilience, and independent journalism.

中文介绍 OpenAI与 AIRPPU 和 WAN-IFRA联合推出AI项目，帮助乌克兰新闻机构增强创新、韧性和独立新闻能力。

Claude proves Fermat 🧮, automated AI researcher 🔬, Z1 efficiency chip ⚡

TLDR AI · 09/06 20:00

中文介绍 该条目暂无具体内容。

An Alien Mind

OpenAI News · 09/06 05:00

Jakub Pachocki reflects on increasingly capable AI and the challenge of keeping it aligned. He calls for stronger safeguards and international coordination.

中文介绍 Jakub Pachocki思考日益强大的AI以及保持其 aligned 的挑战，并呼吁加强安全措施和国际协调。