🔁 LoopKit — build self-improving systems.
PeetPedro
• NEW Articles from Team or Enterprise organizations will get promoted to the main section.
How a 7B Model Beat a 70B Baseline by 46% — Building CodeReviewEnv
lucifer0077
• How to actually evaluate AI agents at scale
darubberduckiee
• • 2Kog Laneformer 2B: The Latency-First Model Behind Kog Inference Engine
kogai
• • 321,567 AI-Designed GID4 Binders: An Open Dataset for Targeted Protein Degradation
hugging-science
• • 1DIFC Regulation 10: The AI Architecture It Demands
Ercin
• Seven Hells
ajsbsd
• Interhuman’s Goblin: “Yeah, Friday at Five”
Interhuman
• • 8Building Moon Bot: A Slack-Native Coding Agent Backed by HuggingFace Buckets
huggingface
• • 50你的 agent 會不會犯傻，常常取決於新聞資料有多新
weileong
• Machine learning for alien climates: Introducing the ThousandWorlds benchmark
hugging-science
• • 4Build real agentic apps using CUGA: two dozen working examples on a lightweight harness
ibm-research
• • 38Agent 不是比較會聊天的 Chatbot：它真正多的是工具、狀態和驗證
weileong
• Anonymize It, Then Attack It: A Tabular Privacy Toolkit on the Hub
t22000t
• Quantizing CLIP to W4A8 with NVIDIA ModelOpt
herooooooooo
• • 3New updates to Papers with Code
nielsr
• NE-MultiSpeech: Building ASR for 8 Underserved Northeast Indian Languages
Badnyal
• • 2V-Zero
hao05
• • 5RAG 準不準不要靠感覺：你需要一份會打臉自己的評估集
weileong
• 從零做一個真的能用的 RAG 知識庫：框架不是第一步，資料清理才是
weileong
• • 1