讀紙 ReadPaper

由 Claude Code Max 翻譯的 AI 論文繁體中文版 · 101 篇

標籤 願望清單
篩選標籤： 開源模型 15 篇 × 清除
2610.01674v1 CC BY 4.0 Google

發明一個資料集：以零種子測量資料集生成能力

Invent a Dataset: Measuring dataset generation abilities with zero seed

Singh, Shivalika、Djurisic, Andrija、Onilude, Gbemileke、Roy, Sudip、Hooker, Sara

LLM開源模型合成資料數據生成數據增強數據集生成零樣本
arXiv 9
Liang-Hsun Huang 許願
2609.13356v1 CC BY 4.0 Alibaba / Qwen Team

ZGCM-1：一個完全開放、極致高效、面向數學與代理式搜尋的基礎模型

ZGCM-1: A Fully Open and Extremely Efficient Foundation Model for Math and Agentic Search

Jiyan He, Guang Liang, Hao Liu, Haoxiang Guan, Jinbo Sun, Junyi Guo, Wenjun Feng, Yantai Xie, Yifei ...

基礎模型開源模型數學推理代理式搜尋混合注意力FP8Muon長上下文
arXiv 15
Liang-Hsun Huang 許願
2608.27370v2 CC BY-SA 4.0 Alibaba / Qwen Team

Puro-2B：窮實驗室在 5090 美元內、用 RTX 5090 訓出的 Qwen2-1.5B

Puro-2B: Poor Lab's Qwen2-1.5B Trained on RTX 5090 within $5090

Kairong Luo、Jiarui Cui、Yaorui Yin、Shengqi Chen 等（清華大學、鵬城實驗室；通訊作者 Kaifeng Lyu、Wenguang Chen）

PretrainingLow-Cost TrainingRTX 5090FP8MuonOpen RecipeScaling Law開源模型
arXiv 22
Liang-Hsun Huang 許願
2506.13585v1 arXiv License Alibaba / Qwen Team

MiniMax-M1：以閃電注意力高效地擴展測試時運算

MiniMax-M1: Scaling Test-Time Compute Efficiently with Lightning Attention

MiniMax（通訊：[email protected]；共 128 位貢獻者，完整名單見原論文附錄 A）

Reasoning ModelLightning AttentionReinforcement LearningCISPOMoELong Context開源模型
arXiv 0
Liang-Hsun Huang 許願
2404.06395v3 CC BY 4.0

MiniCPM：以可擴展的訓練策略揭示小型語言模型的潛力

MiniCPM: Unveiling the Potential of Small Language Models with Scalable Training Strategies

Shengding Hu、Yuge Tu、Xu Han、Chaoqun He、Ganqu Cui、Xiang Long、Zhi Zheng、Yewei Fang、Yuxiang Huang 等（清華大...

Small Language ModelWSD SchedulerScaling LawPre-trainingMoELong Context開源模型
arXiv 23
Liang-Hsun Huang 許願
gh-MoonshotAI-Moonlight MIT License Moonshot AI、UCLA

Muon 可以擴展到 LLM 訓練

Muon is Scalable for LLM Training

Jingyuan Liu、Jianlin Su、Xingcheng Yao、Zhejun Jiang、Guokun Lai、Yulun Du、Yidao Qin、Weixin Xu、Enzhe Lu、...

LLMMuonOptimizerMoEScaling Law開源模型
arXiv 37
Liang-Hsun Huang 許願
gh-MoonshotAI-Kimi-K3 Kimi K3 License Moonshot AI

Kimi K3：開放的前沿智慧

Kimi K3: Open Frontier Intelligence

Kimi Team

LLMMoEKimiAgentLinear AttentionRL開源模型
arXiv 63
Liang-Hsun Huang 許願
2601.07206v1 CC BY 4.0 OpenAI

LLMRouterBench：一個大規模基準與用於 LLM 路由的統一框架

LLMRouterBench: A Massive Benchmark and Unified Framework for LLM Routing

Hao Li, Yiqun Zhang, Zhaoyan Guo, Chenxu Wang, Shengji Tang, Qiaosheng Zhang, Yang Chen, Biqing Qi, ...

LLM RoutingBenchmarkLLMModel EnsemblePerformance-Cost開源模型
arXiv 55
林伯燊 許願
2512.20848v1 CC BY 4.0 Alibaba / Qwen Team

Nemotron 3 Nano：開放高效的 MoE 混合 Mamba-Transformer 智能體推理模型

Nemotron 3 Nano Open Efficient Mixture-of-Experts Hybrid Mamba-Transformer Model for Agentic Reasoning

NVIDIA（300+ 位貢獻者）

LLMMoEMambaHybrid ArchitectureAgenticRLHF開源模型
arXiv 104
林伯燊 許願
2601.20789v1 CC BY 4.0 Alibaba / Qwen Team

SERA：軟驗證的高效程式碼倉庫智能體

SERA Soft-Verified Efficient Repository Agents

Ethan Shen, Daniel Tormoen, Saurabh Shah, Ali Farhadi, Tim Dettmers

LLMCode AgentSWE-benchSynthetic DataSoftware Engineering開源模型
arXiv 101
kerg kerg 許願
2604.12374v1 CC BY 4.0 Alibaba / Qwen Team

Nemotron 3 Super：開源高效混合專家 Mamba-Transformer 模型，面向智能體推理

Nemotron 3 Super Open Efficient Mixture-of-Experts Hybrid Mamba-Transformer Model for Agentic Reasoning

NVIDIA（數百位貢獻者）

LLMMoEMambaAgenticRLHFQuantization開源模型
arXiv 147
Teds Lin 許願
2510.25741v4 CC BY-SA 4.0 ByteDance Seed, UC Santa Cruz, Princeton University, Mila - Quebec AI Institute, University of Montreal, Peking University, CMU, University of Pennsylvania, Conscium, University of Manchester, M-A-P

透過循環語言模型擴展潛在推理 / Ouro

Scaling Latent Reasoning via Looped Language Models

Rui-Jie Zhu, Zixuan Wang, Kai Hua, Tianyu Zhang, Ziniu Li, Haoran Que, Boyi Wei, Zixin Wen, Fan Yin,...

LoopLMLatent ReasoningParameter EfficiencyPre-trainingAdaptive ComputationUniversal TransformerLLM開源模型
arXiv 285
Teds Lin 許願
2512.02498v4 arXiv License MIT

dots.ocr：用單一視覺語言模型完成多語文件版面解析

dots.ocr Multilingual Document Layout Parsing in a Single Vision-Language Model

Yumeng Li, Guang Yang, Hao Liu, Bowen Wang, Colin Zhang (hi lab, Xiaohongshu Inc / 小紅書)

VLMOCRDocument ParsingMultilingualLayout DetectionXiaohongshu開源模型OmniDocBench
arXiv 4
鄭子凡 許願
2501.00656v3 CC BY 4.0 MIT

2 OLMo 2 Furious：完全開放的下一代語言模型家族

2 OLMo 2 Furious

OLMo Team（Pete Walsh, Luca Soldaini, Dirk Groeneveld, Kyle Lo, Shane Arora, Akshita Bhagia, ... Ali ...

LLMOLMo開源模型預訓練Mid-trainingRLVRTüluAllen AI
arXiv 152
林伯燊 許願
2505.09388v1 arXiv License Alibaba / Qwen Team

Qwen3 技術報告

Qwen3 Technical Report

Qwen Team

LLMMoEQwen開源模型推理模型多語言
arXiv 24
Liang-Hsun Huang 許願