新興トピック発見フィード

最終更新: 2026-09-02（UTC基準）／集計窓: 直近約7日

本日のテーマ抽出（AI生成）は未実施のため、直近約7日の「生シグナル」をそのまま表示しています。各項目は一次ソースへのリンク付きです。

スター急増リポジトリ

リポジトリ	勢い
Comfy-Org/ComfyUI	スター増分 +1,037（現在 131,211）
ggml-org/llama.cpp	スター増分 +997（現在 126,814）
open-webui/open-webui	スター増分 +688（現在 150,737）
langgenius/dify	スター増分 +650（現在 154,256）
vllm-project/vllm	スター増分 +642（現在 90,803）
n8n-io/n8n	スター増分 +634（現在 203,171）
unslothai/unsloth	スター増分 +605（現在 75,504）
BerriAI/litellm	スター増分 +504（現在 57,855）
ollama/ollama	スター増分 +475（現在 180,000）
langchain-ai/langchain	スター増分 +441（現在 145,528）
langchain-ai/langgraph	スター増分 +432（現在 40,944）
crewAIInc/crewAI	スター増分 +356（現在 58,014）
openai/whisper	スター増分 +347（現在 108,333）
huggingface/transformers	スター増分 +250（現在 164,728）
danny-avila/LibreChat	スター増分 +247（現在 42,740）
ggml-org/whisper.cpp	スター増分 +180（現在 53,388）
Significant-Gravitas/AutoGPT	スター増分 +176（現在 187,085）
lobehub/lobehub	スター増分 +150（現在 82,174）
modelcontextprotocol/servers	スター増分 +140（現在 90,034）
qdrant/qdrant	スター増分 +137（現在 34,347）

新着 arXiv 論文（cs.AI / cs.CL / cs.LG）

タイトル	カテゴリ	公開日
Beyond Scores: Understanding LLM-as-a-Judge Mechanisms in Summarization Evaluation	cs.CL, cs.LG	2026-09-01
Efficient SWE Agent Benchmarking via Trajectory-Aware Evaluation	cs.SE, cs.AI, cs.CL	2026-09-01
Adaptive Critical Token-Aware Retrieval for Repository-Level Code Generation	cs.SE, cs.AI, cs.CL	2026-09-01
CordisBench: Can Language Models Reason About Component Lifecycles in Dynamic Agent Harnesses?	cs.CL, cs.AI	2026-09-01
The Rise of Verbal Reinforcement Learning	cs.CL, cs.AI	2026-09-01
Facet-0: A Robotic Foundation Model for Contact-Rich Precise Manipulation	cs.RO, cs.LG	2026-09-01
Mechanism Design for Alignment and Control	econ.TH, cs.AI, cs.GT	2026-09-01
StudentSim: Training LLM-based Student Simulators	cs.CL	2026-09-01
Designing Proactive Thought Partners for Writing	cs.HC, cs.AI, cs.CL	2026-09-01
The Structure of Quantization Damage in LLMs: Why the Next Bit Should Be Spent Globally	cs.LG, cs.CL	2026-09-01
Closing Cost-Quality Gap in Document VLMs: Difficulty-Aware Data Curation and Quality-Adjusted Deployment Economics	cs.CL	2026-09-01
Scaling Near-Optimal SFT-RL Annotation Budget Allocation from Small to Large LLMs	cs.CL, cs.AI, cs.LG	2026-09-01
From Production Traffic to Post-Training: Building a Self-Hosted LLM That Covers the Corporate Request Mix	cs.CL	2026-09-01
Selective Agent Guidance via Entropy: Learning Autonomous Policies from Imperfect VLM Teachers	cs.AI, cs.CL, cs.LG	2026-09-01
From Confusion to Clarity: Confusion-Aware Retrieval and Knowledge Injection for Text Classification	cs.CL, cs.AI	2026-09-01
A systematic Approach to constructing a Chance-and-Risk Matrix for Semiconductor Supply Chains	cs.CL	2026-09-01
H3-World: Turning Language Understanding into World Control	cs.CV, cs.AI	2026-09-01
Gradient-Update Mismatch: Rethinking Conflict-Free Training of Physics-Informed Neural Networks	cs.LG	2026-09-01
Retrieved but not ranked: surface-form bias in structural retrieval, from mathematics to agent trajectories	cs.LG, cs.AI, cs.IR	2026-09-01
BS: Take the Hint - Interactive Multitracer PET/CT Lesion Segmentation with a Scribble-Conditioned ResEnc U-Net	cs.CV, cs.AI	2026-09-01

Hacker News 上位スレッド

タイトル	Points
Why your local LLM feels dumber than it is [HN]	507
Nvidia agrees to acquire Hugging Face for $13B [HN]	301
Show HN: We built open OpenRouter that turns usage into a better model [HN]	220
Launch HN: Nori Robotics (YC S26) – A low-cost humanoid robot for development [HN]	194
Movie Scene Map – 13,312 films, series, games, anime and manga [HN]	142
Show HN: Running 104GB Qwen3.8-Flash-Next on 48GB Mac with at ~12 tok/s [HN]	130
Thanks to Lake Ontario, MapQuest is popular all over again [HN]	108
Serve Markdown to AI Agents with Accept Headers [HN]	98
vLLM v0.28.0 [HN]	84
AI Engineer Notebooks – free, framework-free RAG/agents/evals on Colab [HN]	79
Your Open Source Model Could Have a Hidden Time-Release Backdoor [HN]	62
Terminal-Bench-Science: Evaluating AI agents on scientific research workflows [HN]	61
Memory Ordering in CPUs [HN]	51
LLMs are making me lose my savviness [HN]	51
Continuous Diffusion Language Models (CDLM's) [HN]	46
Launch HN: Hebbian Robotics (YC S26) – Build scalable robotics data pipelines [HN]	46
Launch HN: Risklytics (YC S26) – Insurance brokerage for frontier tech companies [HN]	45
Show HN: Watches user sessions, finds bugs that matter, and fixes them [HN]	40
AI Agents and the Refactoring That Never Happens [HN]	40
Show HN: We built the smallest dual-band aircraft tracker [HN]	38

Hugging Face DL数上位

モデル/データ	種別	DL数
sentence-transformers/all-MiniLM-L6-v2	model	256,327,995
google-bert/bert-base-uncased	model	91,982,965
cross-encoder/ms-marco-MiniLM-L6-v2	model	87,411,511
BAAI/bge-small-en-v1.5	model	70,844,329
google/electra-base-discriminator	model	57,108,020
sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2	model	54,423,737
BAAI/bge-m3	model	37,487,482
amazon/chronos-2	model	33,809,631
sentence-transformers/all-mpnet-base-v2	model	24,787,925
lpiccinelli/unidepth-v2-vitl14	model	24,282,648
Qwen/Qwen3-0.6B	model	24,081,897
google-t5/t5-small	model	23,282,469
Comfy-Org/MiniMax-H3	model	22,661,854
FacebookAI/xlm-roberta-base	model	21,313,287
openai/clip-vit-base-patch32	model	20,562,228
timm/mobilenetv3_small_100.lamb_in1k	model	18,815,993
BAAI/bge-reranker-v2-m3	model	18,444,853
trl-internal-testing/tiny-Qwen2ForCausalLM-2.5	model	16,633,375
nomic-ai/nomic-embed-text-v1.5	model	16,632,210
Qwen/Qwen3-8B	model	14,694,928