2026-10-27

On-Policy Self-Training: Data-Free Self-Improvement in Large Language Models

👍 12

10/23 08:00

On-policy self-distillation enables post-training foundation models without relying on stronger external teachers. However, existing approaches rely on ground-truth question-answer pairs, human-curated problem sets, or other supervised signals. In this work, we reformulate on-policy self-distillatio

LoCA: Enabling Multi-Modality in Geospatial Location Encoders

👍 3

10/23 08:00

Location encoders map geographic coordinates to high-dimensional descriptors through contrastive alignment with paired Earth Observation imagery. We prove that this construction is structurally limited: any Lipschitz location encoder produces an output of intrinsic dimension at most 2, regardless of

RSIHub: A Composable Framework for Agent-Evolution Methods

👍 34

10/23 08:00

Agent-evolution methods use execution feedback to improve prompts, skills, tools, and orchestration, but their implementations often rebuild the same experimental machinery. We present RSIHub, an open-source framework for building, comparing, and extending these methods on shared infrastructure. Rec

RynnUnify: Towards Unified Embodied Brain

👍 31

10/23 08:00

We introduce RynnUnify, a 16B unified embodied model that combines strong embodied understanding with long-horizon visual planning through interleaved video--language reasoning. With capabilities spanning spatial reasoning, physical prediction, and grounding, RynnUnify connects an understanding of t

On the Token Collapse Phenomenon in Textual Graph Soft Prompting

👍 0

10/22 08:00

Recent methods for injecting graph structure into LLM prompts include graph soft prompting, in which a graph neural network (GNN) learns graph-specific tokens that capture relational information among entities in textual prompts. G-Retriever combines graph soft prompts with a verbalized description

PaperFormalizer: Author-in-the-Loop Autoformalization of Research Papers in Lean

👍 5

10/22 08:00

Even a carefully proofread paper can leave small gaps in its proofs or its stated assumptions. We introduce PaperFormalizer, a Lean-based agentic pipeline that helps authors find and close such gaps in their own papers and turn each paper into a Lean-verified version. Starting from the LaTeX source,

Speculative Diffusion is a Tale of Two Tradeoffs

👍 3

10/22 08:00

Speculative decoding is a technique that accelerates the inference speed of autoregressive large language models (LLMs) by sampling from a faster, lower quality draft model. Traditionally, a fundamental characteristic of speculative decoding is that it is lossless with respect to the verifier distri

FocusVLA: Future-Intent Guided Dynamic View Routing for Vision-Language-Action Models

👍 1

10/22 08:00

Vision-language-action (VLA) models have become an important paradigm for robotic manipulation by unifying visual perception, language understanding, and action generation. Since reliable manipulation requires sufficient observation of objects, end-effectors, and their spatial relationships, multi-v

MMVistaReason: Toward Open-Data and Post-Training Recipes for Multimodal Reasoning

👍 5

10/22 08:00

Open multimodal reasoning models have benefited from large-scale reasoning supervision, yet reliable post-training remains challenging due to uneven data quality, inefficient supervision construction, imbalanced difficulty, and cross-domain interference. We introduce MMVistaReason (MVR), an open-dat

DraMaze: A Multi-Agent Pipeline with Future-Aware Memory for Diverse Microdrama Script Generation

👍 12

10/22 08:00

The rapid growth of microdramas has increased demand for high-quality script production. Long-form microdrama script generation with LLMs requires both diverse plot progression and long-range narrative coherence. Our observations reveal that candidate continuations increasingly concentrate on fewer

Dual Exam: Controlled Evaluation of Multimodal Agents and World Models

👍 16

10/22 08:00

World Models (WMs) can support Multimodal Agents (MAs) through simulated experience, while Agent trajectories can inform World Model refinement. Evaluating this interaction requires separate evidence about action selection and state evolution. We introduce Dual Exam, which tests Agents under fixed e

HiSTEVE 1.0: Harnessing Human-Interface VLA Agents for Open-World Games

👍 38

10/22 08:00

Building virtual embodied agents that operate in open-ended games like human players is a long-standing goal requiring both long-horizon autonomy and reliable native-interface control. As a procedurally generated open world, Minecraft provides a natural testbed for studying the integration of these

UserArena: Benchmarking Interactive User Simulation for Long-Horizon Shopping Agents

👍 47

10/22 08:00

Agentic recommendation pursues shopper goals through sequences of context-dependent decisions. Developing and evaluating such policies requires faithful recommendation world models that capture both the interactive shopping environment and users' behavioral responses over long interactions. Existing

LLM2Cert: LLM Synthesis of Certificates with SMT/Lean Verification

👍 1

10/21 08:00

Lyapunov functions and barrier certificates give state-space guarantees of stability and safety, but synthesizing and verifying them---especially for nonlinear, high-dimensional systems---remains an open challenge. Existing pipelines verify candidates with numerical solvers (sum-of-squares programs)

TenetBench: A Clause-Anchored Russian Corpus for Evaluating Language Models against Professional Codes of Ethics

👍 12

10/21 08:00

Codes of professional ethics guide and restrict specialists, who increasingly work with assistants built on large language models. Such an assistant can breach the user's code, or help the user breach it, without the user noticing. Evaluations reporting broad categories of harm cannot show how often

Generalized CycleGRPO for Fine-Grained Multimodal Understanding and Grounding

👍 3

10/21 08:00

Fine-grained multimodal intelligence requires models to both understand a designated region and localize it from language. Existing methods learn these directions with separate objectives and task-specific annotations. We present Generalized CycleGRPO, the first unified reinforcement learning framew

TARSphere: A Robot-Oriented Full-Stack Framework for Long-Horizon Embodied Agents in Evolving Physical and Social Environments

👍 8

10/21 08:00

Robots must act under physical and social constraints while adapting to changes in their environment and human needs. Existing embodied-agent systems often focus on specified goals or abstract execution into symbolic actions, limiting evaluation in an evolving world. We introduce TARSphere, a robot-

RoboJudge: Multimodal Language Models as Judges for Embodied Video Generation

👍 3

10/20 08:00

Video generation models are promising foundations for embodied world models, but only if their generated rollouts are physically valid and faithful to task instructions. Multimodal large language models (MLLMs) are increasingly used to judge these properties, yet their agreement with human judgments

DuoLM: Fast and Effective Personalized Recommendation with Language Models

👍 5

10/20 08:00

Language models can make personalized recommendations from rich item content, user history, and feedback. However, evaluating many candidates with full-text inputs incurs substantial online computation and limits serving throughput. We introduce DuoLM, a framework for fast and effective personalized

MultiAgent-OPD: End-to-End Training of Hierarchical Multi-Agent Systems with On-Policy Distillation

👍 8

10/20 08:00

In a hierarchical multi-agent system, a Leader assigns tasks and subagents carry them out. Jointly training the Leader and the subagents is difficult because a final answer gives only indirect feedback to the subagents. An LLM judge can score the performance of the subagents, but these scores depend

Diffusion Language Models Change Speed Across Tasks and Languages

👍 18

10/19 13:47

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 acro

Omni-Woodpecker: Hallucination Correction via Evidence-Prior Rebalancing for Omni LLMs

👍 22

10/19 08:00

Recent Omni LLMs (OLLMs) have demonstrated remarkable capabilities in joint audio-visual reasoning, yet cross-modal hallucinations remain a major challenge. Existing mitigation methods mainly operate at the output level, overlooking the entanglement of grounding evidence and learned priors within fu

How Much Feedback Is Needed to Break Template Protection? The Geometric Vulnerability of Biometric Authentication

👍 1

10/18 08:00

Biometric authentication systems increasingly use privacy-preserving mechanisms to protect sensitive templates. However, the authentication interface itself may still leak information that enables reconstruction attacks. Prior attacks exploit released similarity scores, whereas more advanced systems

Co-adaptive Rule Routing and Constraining for Structured Knowledge Retrieval

👍 1

10/17 08:00

Structured knowledge retrieval acquires task-relevant evidence from structured knowledge sources for downstream tasks. Existing methods rely on unconstrained retrieval or predefined query-rule supervision, such as author-paper rules for scientific discovery and anatomy-drug rules for precision medic

EvalResearchBench: Can AI Agents Build Their Own Feedback for Recursive Self-Improvement?

👍 22

10/16 08:00

Recursive self-improvement (RSI) relies on evaluation feedback to assess progress and guide further research, yet repeatedly running complex benchmarks is costly and slows iteration. Human experts reduce this cost by selecting benchmark subsets or designing compact suites. We ask whether AI agents c

Kolab: Benchmarking Human-Agent Collaboration in Long-Horizon Workflows

👍 5

10/15 08:00

When humans collaborate with AI agents on long-horizon tasks, what happens to the agents' contributions? For solving long-horizon tasks, humans place greater importance on ongoing interaction with AI agents (e.g., revising plans) than on an agent's ability to complete the task in a single attempt. H

The LLM-Attacker: A Unified Red-Team Framework for Privacy Evaluation of Split and Distributed LLM Systems

👍 10

10/15 08:00

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

JockeyBench: Evaluating LLM Agents on Horse-Racing Decision-Making

👍 2

10/14 08:00

Repeated event forecasting requires extracting predictive signals from noisy evidence and evaluating their value across successive decisions. We introduce JockeyBench, a benchmark and evaluation protocol that studies this problem through historical Japanese horse racing. The protocol combines time-g

SliderEdit++: Improved Slider Trajectories in Instruction-Based Image Editing

👍 1

10/14 08:00

Slider-based control aims to give users continuous, fine-grained control over the intensity of individual edits in instruction-based image editing. However, existing approaches fail to deliver reliable continuous controllability: traversing a slider can produce abrupt visual jumps and non-monotonic

What Git Does Not Record: Provenance Legibility in Two AI-Assisted Research Repositories

👍 1

10/09 08:00

Publishers and funders increasingly ask authors to disclose how AI tools contributed to a piece of research. In computational work the repository is where an author or an auditor would look for the answer, and we ask what it can actually supply. We audit two single-investigator repositories by the s

How to Make Animated Short Videos to 10X Your App Revenue

@tobiadonadon_ · 5.7K 粉丝 · 139.3K 阅 · 513 赞 · 44 转

10/26 03:55

making an animated short normally means an animator, footage, or hours inside an editing tool and if you just have a verse, a quote or a small story in your head, that's a lot of work for something

中文介绍 博主分享如何用低成本制作动画短视频，以提升应用收入。

Replit and the dawn of the self-driving company

@pmarca · 6.7M 粉丝 · 92.9K 阅 · 587 赞 · 78 转

10/27 00:08

One thing I love about Amjad is, sometimes he really says the thing and articulates something that other people weren’t brave enough or clear enough in their thinking to express. He has one of those

中文介绍 博主探讨 Replit 在自驱动公司发展中的作用。

How the Claude Code team does code review

@addyosmani · 465.0K 粉丝 · 70.4K 阅 · 1.2K 赞 · 85 转

10/26 14:27

We previously mentioned that Claude writes about 90 percent of the code for Claude Code. So how do we do code review at this scale? Who reads code anyway? What happens if code goes wrong? This post

中文介绍 博主揭秘 Claude Code 团队如何进行大规模代码审查。

How to make an AI UGC video ad, a step by step guide

@SayItAgainAI · 931 粉丝 · 51.6K 阅 · 502 赞 · 62 转

10/26 06:13

An AI UGC video is a short vertical video in which a generated presenter films themselves with your product and talks about it, the way a person would on their phone. Nobody is hired, nothing is

中文介绍 博主提供制作 AI 用户生成内容视频广告的详细步骤指南。

The five components of building a "company harness"

@hwchase17 · 136.6K 粉丝 · 39.2K 阅 · 521 赞 · 60 转

10/25 01:08

How do you build a company harness? We've built internal agents at LangChain across marketing, sales, and engineering, and they have completely changed how those teams work. They weren't easy to get

中文介绍 博主分享构建“公司 harness”的五个要素。

How I work with Agents

@thorstenball · 57.2K 粉丝 · 35.8K 阅 · 831 赞 · 59 转

10/25 23:58

... as of October 2026. Use the latest and greatest models. Everything else is potentially a waste of time. If a model is wrong 10% of the time, you don't know which 10%, so you end up making multiple

中文介绍 博主阐述如何使用最新 AI 模型及应对模型错误的方法。

Will the real Anthropic please stand up

MIT Tech Review AI · 10/27 01:45

This story originally appeared in The Algorithm, our weekly newsletter on AI. To get stories like this in your inbox first, sign up here. After weeks of frenzied speculation, Anthropic has filed its IPO prospectus. The 335-page document (100 pages or so are focused on risks) lays out the company’s u

中文介绍 Anthropic提交IPO招股说明书，文件长335页，其中约100页涉及风险。

GPT-6.1 Sol helps CodeRabbit strengthen security across complex codebases

OpenAI News · 10/26 21:45

CodeRabbit uses GPT-6.1 Sol to investigate security risks across complex codebases. Its security agent helps developers review findings and proposed fixes.

中文介绍 CodeRabbit利用GPT-6.1 Sol调查复杂代码库中的安全风险，其安全代理帮助开发者审查发现和提议的修复方案。

ChatGPT’s expanding world of work

OpenAI News · 10/26 18:00

Explore OpenAI Signals data on 165 work activities across nearly 100 countries, revealing how people use ChatGPT for work and how usage changes over time.

中文介绍 OpenAI数据显示，ChatGPT在近100个国家用于165项工作活动，使用情况随时间变化。

The focus on AI safety disasters ignores everyday harms

MIT Tech Review AI · 10/26 17:01

Consider what passes for success in technology policy. An AI company promises that its chatbot will not encourage suicide. A social media platform makes it more difficult for children to buy fentanyl. These are important protections, and they should have existed all along. But companies can clear th

中文介绍 关注AI安全灾难忽视了日常伤害，例如技术政策中AI公司承诺其聊天机器人不会鼓励自杀。

How Fanatics and OpenAI streamline the fan experience

OpenAI News · 10/26 15:00

Fanatics uses OpenAI to help fans complete their game-day outfits with Shop the Look. Testing across roughly 30,000 MLB products showed 10% higher engagement.

中文介绍 Fanatics使用OpenAI帮助球迷完成比赛日装扮，测试显示参与度提高10%。

How Ultrafast cut incident investigation from hours to minutes at OpenAI

OpenAI News · 10/26 15:00

See how OpenAI engineers use Codex Ultrafast to investigate incidents in minutes. Explore faster security analysis, code refactoring, and daily workflows.

中文介绍 OpenAI工程师使用Codex Ultrafast将事件调查时间从数小时缩短至数分钟。

Working with academic communities to build scientific understanding

OpenAI News · 10/26 08:00

OpenAI shares model results on the Kakeya conjecture, Exponential Time Hypothesis, and quantum chromodynamics—and how researchers can verify and build on them.

中文介绍 OpenAI与学术社区合作，分享模型结果，包括Kakeya猜想、指数时间假设和量子色动力学。

Introducing GPT-6.1 Terra

OpenAI News · 10/26 08:00

Bringing fast and capable computer use to everyone

中文介绍 OpenAI推出GPT-6.1 Terra，提供快速且强大的计算机使用体验。

Apple’s 20B model 📱, Kimi K3.1 Pro 🧠, Block Dessa 🗣️, frontier training security 🔐

TLDR AI · 10/26 08:00

中文介绍 报道了苹果的20B模型、Kimi K3.1 Pro、Block Dessa以及前沿训练安全。

[AINews] Kimi 3.1 Pro 104BA-2.8T: Fable-5.1-Class Open Weights narrows Closed-AI lead to 1 month... and Pro Highspeed API with 2-3× speed

Latent Space · 10/24 15:18

We have Fable at home... or so we are being told by paid shills

中文介绍 Kimi 3.1 Pro 104BA-2.8T：Fable-5.1级开放重量缩小了Closed-AI领先优势至1个月，并推出2-3倍速度的Pro Highspeed API。

Rapidly scaling online storage to serve over 1 billion ChatGPT users (Part II)

OpenAI News · 10/24 01:00

Learn how OpenAI scaled Habitat with reliable multi-tenancy, global caching, and Azure Cosmos DB to deliver 12× faster reads for ChatGPT.

中文介绍 OpenAI通过可靠的共享租户、全局缓存和Azure Cosmos DB将Habitat扩展至支持超过10亿ChatGPT用户，读取速度提高12倍。

LitPivot: Developing research ideas through dynamic contextualization within the literature landscape

Hugging Face Blog · 10/24 00:15

中文介绍 通过文献景观中的动态语境化开发研究想法。

AI Is Getting Good at Math. Here’s Why.

Latent Space · 10/23 23:19

AI models are solving more mathematical problems. A pattern is emerging in their successes that could help us predict which problems they’ll solve next.

中文介绍 AI在解决数学问题方面表现出色，出现了一种成功模式，可能有助于预测它们将解决哪些问题。

[AINews] Black October

Latent Space · 10/23 18:09

the Cyber onslaught continues.

中文介绍 网络攻击持续进行中。

Securing our frontier model training environment

OpenAI News · 10/23 18:00

OpenAI explains how we strengthened containment for frontier model training and evaluation, including offline environments and high-compute red-teaming.

中文介绍 OpenAI解释了如何加强前沿模型训练和评估的隔离，包括离线和高度计算的红队演练。