数据来源：ArXiv Domain LLM Domain Papers1. The FlightFinder Occurrence Corpus: harmonising 128 aviation-safety source codes into one deduplicated cross-national accident datasetAbstract:National aviation-investigation databases are deep but territorial; global compilations such as the Aviation Safety Network and the Bureau of Aircraft Accidents Archives are hand-curated, apply their own inclusion thresholds, and publish no per-record lineage. The FlightFinder Occurrence Corpus harmonizes 279,485 oc ...

数据来源：ArXiv Domain LLM Domain Papers1. The FlightFinder Occurrence Corpus: harmonising 128 aviation-safety source codes into one deduplicated cross-national accident datasetAbstract:National aviation-investigation databases are deep but territorial; global compilations such as the Aviation Safety Network and the Bureau of Aircraft Accidents Archives are hand-curated, apply their own inclusion thresholds, and publish no per-record lineage. The FlightFinder Occurrence Corpus harmonizes 279,485 oc ...

数据来源：ArXiv Domain LLM Domain Papers1. The FlightFinder Occurrence Corpus: harmonising 128 aviation-safety source codes into one deduplicated cross-national accident datasetAbstract:National aviation-investigation databases are deep but territorial; global compilations such as the Aviation Safety Network and the Bureau of Aircraft Accidents Archives are hand-curated, apply their own inclusion thresholds, and publish no per-record lineage. The FlightFinder Occurrence Corpus harmonizes 279,485 oc ...

数据来源：ArXiv Domain LLM Domain Papers1. Li$^{2}$SAO: Lifting Light Singl-Rollout Asynchronous RL Optimization via Group-Relative Actor-Critic BootstrappingAbstract:Asynchronous reinforcement learning (RL) is becoming a mainstream paradigm for post-training large language models (LLMs), as it improves training efficiency by updating the model as rollouts arrive. Recent work argues that group-wise sampling in GRPO does not naturally fit asynchronous RL training because rollout lengths follow a lon ...

数据来源：ArXiv Domain LLM Domain Papers1. From Fragmentation to Fusion: A Survey and Roadmap for Omnimodal EmbeddingsAbstract:Embedding models underpin virtually every modern AI system, from search engines to recommendation systems. Despite shared mathematical foundations, research on embedding models remains fragmented by modality: the communities develop largely in isolation, seldom exchanging insights. We argue that this fragmentation obscures a deeper convergence. Driven by contrastive learnin ...

数据来源：ArXiv Domain LLM Domain Papers1. Rethinking Next-Visit Diagnosis Prediction Beyond Aggregate Metrics: Retrieval—Memory Multi-Agent AdjudicationAbstract:Next-visit diagnosis prediction forecasts the diagnoses documented at a patient’s next encounter, turning retrospective electronic health records into prospective clinical prediction. However, conventional aggregate evaluation obscures its multifaceted nature: distributionally, models must handle variable-size diagnosis sets and long-taile ...

数据来源：ArXiv Domain LLM Domain Papers1. Discourse Analysis of Legislative Bill DiscussionsAbstract:Legislative hearings are distinct communication events with their own human recognizable structure and logic, yet they have been analyzed as distinct phenomena beyond their transcripts. Understanding legislative hearings at a deeper level is a major step in analyzing the work and power relationships within democratic bodies. In this work, we present a novel end-to-end pipeline for parsing rhetorica ...

数据来源：github.com/trending global LanguagesNVIDIA/OpenShellOpenShell is the safe, private runtime for autonomous AI agents. ⭐ Stars: 12795 🍴 Forks: 0 📝 Language: Rust debpalash/VoiceStudioVoiceStudio is the open-source, fully-local ElevenLabs alternative — voice cloning, voice design, video dubbing, dictation, transcription & audiobook creation in 646 languages. ⭐ Stars: 50523 🍴 Forks: 0 📝 Language: Python mvschwarz/openrigMulti-agent harness that runs Claude Code and Codex togethe ...

数据来源：github.com/trending global LanguagesDietrichGebert/ponytailMakes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote. ⭐ Stars: 150595 🍴 Forks: 0 📝 Language: JavaScript mattpocock/skillsSkills for Real Engineers. Straight from my .agents directory. ⭐ Stars: 273957 🍴 Forks: 0 📝 Language: Shell NVIDIA/OpenShellOpenShell is the safe, private runtime for autonomous AI agents. ⭐ Stars: 14052 🍴 Forks: 0 📝 Language: Rust firebase/f ...

数据来源：github.com/trending global LanguagesPanniantong/Agent-ReachGive your AI agent eyes to see the entire internet. Read & search Twitter, Reddit, YouTube, GitHub, Bilibili, XiaoHongShu — one CLI, zero API fees. ⭐ Stars: 88715 🍴 Forks: 0 📝 Language: Python JuliusBrussee/caveman🪨 why use many token when few token do trick. Viral skill + proxy for coding agents that cuts 65% of tokens by talking like a caveman. ⭐ Stars: 109122 🍴 Forks: 0 📝 Language: Go obra/superpowersAn agentic s ...

数据来源：github.com/trending global LanguagesDietrichGebert/ponytailMakes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote. ⭐ Stars: 153522 🍴 Forks: 0 📝 Language: JavaScript pbakaus/impeccableThe design language that makes your AI harness better at design. ⭐ Stars: 75383 🍴 Forks: 0 📝 Language: JavaScript affaan-m/ECCThe agent harness performance optimization system. Skills, instincts, memory, security, and research-first development f ...

数据来源：github.com/trending global Languagestester-army/e2eNext generation e2e testing framework for web and mobile apps. ⭐ Stars: 3202 🍴 Forks: 0 📝 Language: TypeScript pbakaus/impeccableThe design language that makes your AI harness better at design. ⭐ Stars: 76355 🍴 Forks: 0 📝 Language: JavaScript coreyhaines31/marketingskillsMarketing skills for Claude Code and AI agents. CRO, copywriting, SEO, analytics, and growth engineering. ⭐ Stars: 53111 🍴 Forks: 0 📝 Language: JavaScript ...

数据来源：github.com/trending global Languagestester-army/e2eNext generation e2e testing framework for web and mobile apps. ⭐ Stars: 4935 🍴 Forks: 0 📝 Language: TypeScript thedotmack/claude-memPersistent Context Across Sessions for Every Agent – Captures everything your agent does during sessions, compresses it with AI, and injects relevant context back into future sessions. Works with Claude Code, OpenClaw, Codex, Gemini, Hermes, Copilot, OpenCode + More ⭐ Stars: 96677 🍴 Forks: 0 📝 Languag ...

数据来源：github.com/trending global Languagestester-army/e2eNext generation e2e testing framework for web and mobile apps. ⭐ Stars: 6428 🍴 Forks: 0 📝 Language: TypeScript mattpocock/skillsSkills for Real Engineers. Straight from my .agents directory. ⭐ Stars: 278225 🍴 Forks: 0 📝 Language: Shell earthtojake/text-to-cadGive your agent CAD superpowers. ⭐ Stars: 18024 🍴 Forks: 0 📝 Language: Python boykopovar/AnyPS5Tool for automatic PS5 executables porting to Linux and Windows ⭐ Stars: ...

数据来源：github.com/trending global Languagesmorluto/reaReverse engineer anything with agents, from app behavior down to native binaries. ⭐ Stars: 15934 🍴 Forks: 0 📝 Language: TypeScript mattpocock/skillsSkills for Real Engineers. Straight from my .agents directory. ⭐ Stars: 279760 🍴 Forks: 0 📝 Language: Shell boykopovar/AnyPS5Tool for automatic PS5 executables porting to Linux and Windows ⭐ Stars: 10893 🍴 Forks: 0 📝 Language: C++ ayghri/i-have-adhdA skill to stop your coding agent ...