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： Agent 12 篇 × 清除
2609.24974v1 arXiv License Google

Harness-Zero：透過代理人即Harness進行Harness蒸餾

Harness-Zero: Harness Distillation via Agent-as-Harness

Ye, Haoran、Lu, Yuxing、Dong, Haonan、Su, Zhaochen、Song, Guojie

LLMKnowledge DistillationAgentHarnessLLM-as-JudgeSynthetic DataSelf-Evolution
arXiv 1
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
2608.24571v1 CC BY 4.0 Appier AI Research、National Taiwan University

大型語言模型代理人的工具創建與工具使用聯合最佳化

Joint Optimization of Tool Creation and Use for Large Language Model Agents

Zhi Rui Tam、Chieh-Yen Lin、Yun-Nung Chen、Shao-Hua Sun、Hung-yi Lee

LLMAgentTool UseRLVRDAPOReward Design工具創建
arXiv 70
Liang-Hsun Huang 許願
2603.24621v2 arXiv License MIT

ARC-AGI-3：前沿代理式智慧的全新挑戰

ARC-AGI-3 A New Challenge for Frontier Agentic Intelligence

ARC Prize Foundation

ARC-AGIAGIBenchmarkAgentReasoning互動式環境評測
arXiv 1
Liang-Hsun Huang 許願
2606.25996v2 arXiv License Meta

Autodata：以代理人扮演資料科學家來生成高品質合成資料

Autodata: An agentic data scientist to create high quality synthetic data

Ilia Kulikov, Chenxi Whitehouse, Tianhao Wu, Yixin Nie, Swarnadeep Saha, Eryk Helenowski, Weizhe Yua...

合成資料AgentSelf-InstructGRPOLLM-as-a-Judge推理Meta-Optimization
arXiv 1
kerg kerg 許願
2605.14038v2 CC BY 4.0 Meta

模型自適應的工具必要性揭示了 LLM 工具使用中的「知行落差」

Model-Adaptive Tool Necessity Reveals the Knowing-Doing Gap in LLM Tool Use

Yize Cheng, Chenrui Fan, Mahdi JafariRaviz, Keivan Rezaei, Soheil Feizi（University of Maryland, Coll...

LLMTool UseAgentProbingMeta-cognitionInterpretability知行落差
arXiv 56
Teds Lin 許願
2604.09791v1 CC BY 4.0

Pioneer Agent：生產環境中小型語言模型的持續改進

Pioneer Agent Continual Improvement of Small Language Models in Production

Dhruv Atreja, Julia White, Nikhil Nayak, Kelton Zhang, Henrijs Princis, George Hurn-Maloney, Ash Lew...

Small Language ModelAgentContinual LearningProductionMonte Carlo Graph SearchAutoMLFine-tuning
arXiv 156
Liang-Hsun Huang 許願
2604.15034v2 CC BY 4.0 Google

Autogenesis：自我演化的智能體協定

Autogenesis A Self-Evolving Agent Protocol

Wentao Zhang, Zhe Zhao, Haibin Wen, Yingcheng Wu, Ming Yin, Bo An, Mengdi Wang

AgentSelf-EvolutionProtocolLLMMulti-AgentTool UseMCP
arXiv 58
Teds Lin 許願
2604.15597v1 arXiv License

當你委派工作時，LLMs 會污染你的文件

LLMs Corrupt Your Documents When You Delegate

Philippe Laban, Tobias Schnabel, Jennifer Neville

LLMDocument EditingHallucinationAgentBenchmarkLong-Horizon EvaluationDelegated Work
arXiv 0
Liang-Hsun Huang 許願
2604.18131v1 CC BY 4.0

訓練 LLM 智能體透過世界知識探索實現自發、無獎勵的自我演化

Training LLM Agents for Spontaneous Reward-Free Self-Evolution via World Knowledge Exploration

Qifan Zhang, Dongyang Ma, Tianqing Fang, Jia Li, Jing Tang, Nuo Chen, Haitao Mi, Yan Wang

LLMAgentSelf-EvolutionReinforcement LearningWeb Agent世界知識
arXiv 101
Teds Lin 許願
2210.03629v3 CC BY 4.0 ¹普林斯頓大學 (Princeton University) 計算機科學系；²Google Research, Brain team

ReAct：在語言模型中協同推理與行動

ReAct Synergizing Reasoning and Acting in Language Models

Shunyu Yao¹*、Jeffrey Zhao²、Dian Yu²、Nan Du²、Izhak Shafran²、Karthik Narasimhan¹、Yuan Cao²

LLMAgentReasoningTool UsePromptingCoTICLR2023
arXiv 99
Thomas Liang 許願
2604.08545v1 arXiv License ¹Accio Team, Alibaba Group；²華中科技大學 (Huazhong University of Science and Technology)

智慧行動：在代理人式多模態模型中培養後設認知的工具使用

Act Wisely Cultivating Meta-Cognitive Tool Use in Agentic Multimodal Models

Shilin Yan¹*†‡、Jintao Tong¹,²*、Hongwei Xue¹†、Xiaojun Tang¹、Yangyang Wang¹、Kunyu Shi¹、Guannan Zhang¹、...

MultimodalAgentTool UseRLHFGRPOHDPOMeta-CognitionVLM
arXiv 2
Teds Lin 許願