2026年8月13日 技术热点总结

项目简介：Strip multi-vendor AI provenance marks: Unicode text hygiene, statistical rewrite hooks, and C2PA/metadata from PNG/JPEG/SVG/PDF/DOCX/HTML/MD

技术栈：Python、agent-skill、ai、c2pa、claude、provenance、synthid、watermark

核心介绍：_ _ _ ____ ___ ____ ____ _ _ ____ ____ _ _ ____ ____ ____ _ _ ____ _ _ ____ ____ Agent skill + stdlib Python scripts to strip multi-vendor AI provenance marks from text and files — for privacy and hygiene on content you own. Vendors / ecosystems (class-level): Claude, Gemini / SynthID-Text, OpenAI provenance surfaces, open-LLM Kirc…

项目简介：飞鼠格式 FlyingMouse Format – Windows 万能文件格式转换工具（离线可用，内置 FFmpeg/LibreOffice/Poppler/Tesseract）。图片/文档/表格/PPT/PDF/音视频/WPS 格式互转 + OCR + 批量转换

技术栈：JavaScript、desktop-app、electron、ffmpeg、file-converter、format-converter、libreoffice、ocr、pdf

核心介绍：> A mouse-themed, offline Windows file converter. / 一款鼠鼠主题、可离线使用的 Windows 文件格式转换工具。 下载最新版 / Download · 问题反馈 / Issues > NCM 说明：仅保证支持 music.163.com 对应客户端下载的音乐文件。其他网站或来源虽然扩展名也可能是 .ncm，但内部格式不同，不属于本项目的兼容范围。 1. 在 Releases 下载 v0.3.5 对应系统的安装包。 2. 安装并启动 FlyingMouse Format。 3. 拖入文件，选择目标格式并开始转换。 4. 选择保存位置；软件会记住目标格式与保存目录。 > 源码仓库不包含体积较大的 FFmpeg、LibreOffic…

技术栈：Shell

核心介绍：src=”https://github.com/gvzdv/claudish-to-english/releases/download/assets/comparison.png” alt=”Side-by-side comparison: a dense, jargon-heavy Claude message labeled ‘Claudish’ on the left, and its plain-English rewrite on the right”> A Claude Code plugin that shows a plain-English rewrite of each assistant message, produced by a local LLM via ollama. It is display-only: Claude’s own reasoning and the saved transcript keep the original text — only what you read on screen changes.

项目数据：⭐ 616 Stars，🍴 37 Forks

项目简介：A minimal chatbot template built with Next.js, AI SDK, shadcn/ui, shadcn/react, shadcn/typeset. It runs on the Vercel AI Gateway.

技术栈：TypeScript

核心介绍：A minimal chatbot template built with Next.js, the AI SDK, shadcn/ui, shadcn/react, shadcn/typeset and the Vercel AI Gateway. That’s it — no configuration needed. Vercel deployments authenticate to the AI Gateway automatically via OIDC, and usage runs on your team’s AI Gateway credits. Then give the app a gateway credential, either by pulling an OIDC token from your linked Vercel project:

关键特性：Streaming chat with mar…

Recent optimization studies on Retrieval-Augmented Generation (RAG) have exploited chunk-level KV cache reuse to avoid processing long retrieved contexts for higher efficiency, while significant information redundancy and noise still remain in the coarse-grained chunks. This paper optimizes the Pareto frontier under low prefill latency constraints while maximizing accuracy by proposing CoinRAG (Contextualized Information Nugget KV Cache Reuse for Long-Context RAG). The name metaphorically reflects our core mechanism: much like assembling small tokens (or "coins") to accumulate a larger valu…

Long-document understanding often requires reasoning over many visually rich pages, making inference costly and prone to context rot. In this work, we propose InSight-doc, an agentic visual perception framework that treats visual resolution as an adaptive reasoning-time resource. InSight-doc starts from low resolution and selectively zooms into high-resolution regions for finer evidence, without relying on any external retriever. To train such an agent, we construct an active-perception corpus of 17.9K high-quality SFT examples with region-level zoom-in trajectories, accompanied by 19.2K ha…

The Large Language Model from Power Law Decoder Representations (PLDR-LLM) and its attention, Power Law Graph Attention (PLGA), replace the fixed bilinear form of scaled dot-product attention (SDPA) with a learned, input-generated bilinear operator G_{LM}, built from a positive tensor A_{LM} by elementwise power laws. The architecture is fully specified, verified against pinned reference releases; claims are labeled theorem, conditional theorem, measurement, or conjecture. Unconditionally: PLGA contains SDPA exactly at G_{LM}=I; A_{LM} and A_P are strictly entrywise positive, with Perron-Fr…

We present 360CityArena, a benchmark for evaluating the urban exploration capabilities of embodied agents within a photorealistic environment constructed from 360-degree videos. Existing outdoor benchmarks either lack sufficient photorealism or complexity, resulting in a considerable gap from real-world urban environments. 360CityArena is built on a realistic reconstruction of the Akihabara district in Tokyo, Japan, using 602 360-degree video segments covering 85 streets, and consists of 175 meticulously human-crafted tasks. It encompasses three task categories: Environment Understanding, P…

Large language model evaluations typically focus on performance under nominal conditions, creating an illusion of capability where models comfortably walk a narrow, highly optimized generation corridor. In real-world deployments, however, complex system prompts, safety guardrails, and structural constraints continuously force models off this nominal path, driving a divergence between benchmark scores and deployment performance. To address this issue, we introduce Decoding-Level Taboo, a zero-prompt diagnostic stress test that intervenes directly in logit space at runtime, forcing models out…

Sparse mixture-of-experts (MoE) layers expand recommendation capacity through conditional computation, yet a trained checkpoint still stores and routes over its full expert bank. We study a deployment problem: convert that checkpoint to a smaller standard MoE under an explicit expert budget, without adding a compression-specific online module. To address this, we introduce UniMoMo, a post-training compression framework formulated as a constrained graph coarsening problem. Rather than relying on parameter distance, UniMoMo groups experts based on their functional similarity, using an unlabel…