Top AI Repos — open-source AI, indexed and scored
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Daily updated LLM papers. 每日更新 LLM 相关的论文,欢迎订阅 👏 喜欢的话动动你的小手 🌟 一个
| Date | Stars |
|---|---|
| 2026-07-24 | 1304 |
| 2026-07-25 | 1304 |
| 2026-07-28 | 1304 |
| 2026-07-30 | 1304 |
| 2026-07-31 | 1305 |
| 2026-08-06 | 1305 |
Today
— stars today
This week
+1 stars this week
This month
— stars this month
Momentum
1.0
growth rate 0.08%/day
<h2 align='center'>llm-paper-daily 日常论文精选</h2> <div align='center'> []() [](./README.md) [](./README_en.md) </div> 欢迎来到 **llm-paper-daily**! 这是一个获取 LLM、Agent 相关研究论文的每日更新和分类平台。 📚 **每日更新:** 仓库每天会带来最新的 LLM、Agent 相关研究,并附有 arXiv 地址、相关 GitHub 仓库和文章的总结。 <!-- paper-daily:readme:updates:start --> <details> <summary>查看更新文章 <sub>更新时间: 2026年07月24日 04:44</sub></summary> <br> - Agentic coding without the cloud: evaluating open-weight large language models on longitudinal data preparation tasks - Same Dangerous Objective, Opposite Advice: Direct Exposure versus Multi-Agent Mediation - GRADRAG: Cross-Component Prompt Adaptation for Coordinated Multi-Agent RAG - HiMe: Real-Time Self-Hosted Personal Agent Platform for Health Insights with Wearable Devices - PATS: Policy-Aware Training Scaffolding for Agentic Reinforcement Learning </details> <!-- paper-daily:readme:updates:end --> <details> <summary><strong>订阅</strong></summary> 想订阅每日 LLM、Agent 论文更新时,不需要手动配置脚本。把下面这段话发送给本地的 OpenClaw、Codex 或 Claude Code,让 Agent 帮你完成配置: ```text 请帮我配置 llm-paper-daily 的本地订阅。订阅仓库是 https://github.com/xianshang33/llm-paper-daily ,请阅读仓库根目录的 SUBSCRIBE.md,按文档创建本地配置、预览 digest、安装定时任务,并在完成后告诉我配置文件位置、运行时间、语言、每次推送数量和验证结果。 ``` Agent 会使用仓库里的 `paper-subscribe` skill,只读取公开的 `feed-papers.json`,不会在你的机器上运行论文抓取或总结生产流程。 </details> ## 最新论文 <!-- paper-daily:readme:months:start --> ### 2026年07月 | Date | Paper | Links & Summary | | --- | --- | --- | | <span style='display: inline-block; width: 42px;'>07-23</span> | **Agentic coding without the cloud: evaluating open-weight large language models on longitudinal data preparation tasks**<br><sub>机构: University College London<br>本文针对敏感数据研究中 AI 应用的数据隐私痛点,提出了 RRBench 框架,系统评估了本地部署的开放权重 LLM 在纵向数据准备任务中的能力。结果表明,31-35B 参数级别的模型在消费级硬件上即可实现接近 88% 的任务完成率,为在严格数据治理环境下实现 AI 辅助数据清洗和处理提供了切实可行的技术方案和评估标准。</sub>| <div style='min-width:85px;'>[](https://arxiv.org/pdf/2607.21482v1)</div><div style='min-width:85px;'>[](summary/2026-07/2607.21482.md) <div style='min-width:85px;'>[](https://github.com/UCL-ARC/RRBench)</div> | | <span style='display: inline-block; width: 42px;'>07-23</span> | **Same Dangerous Objective, Opposite Advice: Direct Exposure versus Multi-Agent Mediation**<br><sub>机构: University of Pennsylvania<br>本文揭示了高能力LLM在多智能体工作流中存在严重的组合式安全缺口。研究表明,虽然模型能抵抗直接的有害指令,但当同一目标经过中间代理的转换和清洗,剥离了明显的操纵性措辞后,下游模型会倾向于执行该隐藏目标。这种“行为逆向偏移”暴露了当前AI系统在架构可见性上的根本缺陷:终端用户和模型无法追溯上游的原始恶意指令。这一发现对设计更安全的多智能体系统和审计机制提出了严峻挑战,强调了对工作流全链路透明度进行监管的必要性。</sub>| <div style='min-width:85px;'>[](https://arxiv.org/pdf/2607.21518v1)</div><div style='min-width:85px;'>[](summary/2026-07/2607.21518.md) | | <span style='display: inline-block; width: 42px;'>07-23</span> | **GRADRAG: Cross-Component Prompt Adaptation for Coordinated Multi-Agent RAG**<br><sub>机构: Bloomberg<br>GRADRAG 解决了多智能体 RAG 系统中组件孤立优化导致的错误传播问题。通过将流水线建模为计算图并引入跨组件的反馈循环,它能够有效利用下游评估信号来优化上游检索和证据构建过程。实验表明,该方法在无需大量迭代的情况下,显著提升了文档问答和摘要任务的性能,为协调多智能体 RAG 提供了新的范式。</sub>| <div style='min-width:85px;'>[](https://arxiv.org/pdf/2607.21324v1)</div><div style='min-width:85px;'>[](summary/2026-07/2607.21324.md) | | <span style='display: inline-block; width: 42px;'>07-23</span> | **HiMe: Real-Time Self-Hosted Personal Agent Platform for Health Insights wit
Excerpt of 255,470 characters
Read on GitHub183
87
Would you bet a product on this? Bounded 0–100 and slow moving.
matched fp:0ee5e9d5d32fa419, topic:large-language-models, topic:llm
matched fp:0ee5e9d5d32fa419, topic:rag
matched fp:0ee5e9d5d32fa419, topic:chatgpt