[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"similar-pat-jj--Awesome-Adaptation-of-Agentic-AI":3,"tool-pat-jj--Awesome-Adaptation-of-Agentic-AI":62},[4,18,26,36,46,54],{"id":5,"name":6,"github_repo":7,"description_zh":8,"stars":9,"difficulty_score":10,"last_commit_at":11,"category_tags":12,"status":17},4358,"openclaw","openclaw\u002Fopenclaw","OpenClaw 是一款专为个人打造的本地化 AI 助手，旨在让你在自己的设备上拥有完全可控的智能伙伴。它打破了传统 AI 助手局限于特定网页或应用的束缚，能够直接接入你日常使用的各类通讯渠道，包括微信、WhatsApp、Telegram、Discord、iMessage 等数十种平台。无论你在哪个聊天软件中发送消息，OpenClaw 都能即时响应，甚至支持在 macOS、iOS 和 Android 设备上进行语音交互，并提供实时的画布渲染功能供你操控。\n\n这款工具主要解决了用户对数据隐私、响应速度以及“始终在线”体验的需求。通过将 AI 部署在本地，用户无需依赖云端服务即可享受快速、私密的智能辅助，真正实现了“你的数据，你做主”。其独特的技术亮点在于强大的网关架构，将控制平面与核心助手分离，确保跨平台通信的流畅性与扩展性。\n\nOpenClaw 非常适合希望构建个性化工作流的技术爱好者、开发者，以及注重隐私保护且不愿被单一生态绑定的普通用户。只要具备基础的终端操作能力（支持 macOS、Linux 及 Windows WSL2），即可通过简单的命令行引导完成部署。如果你渴望拥有一个懂你",349277,3,"2026-04-06T06:32:30",[13,14,15,16],"Agent","开发框架","图像","数据工具","ready",{"id":19,"name":20,"github_repo":21,"description_zh":22,"stars":23,"difficulty_score":10,"last_commit_at":24,"category_tags":25,"status":17},3808,"stable-diffusion-webui","AUTOMATIC1111\u002Fstable-diffusion-webui","stable-diffusion-webui 是一个基于 Gradio 构建的网页版操作界面，旨在让用户能够轻松地在本地运行和使用强大的 Stable Diffusion 图像生成模型。它解决了原始模型依赖命令行、操作门槛高且功能分散的痛点，将复杂的 AI 绘图流程整合进一个直观易用的图形化平台。\n\n无论是希望快速上手的普通创作者、需要精细控制画面细节的设计师，还是想要深入探索模型潜力的开发者与研究人员，都能从中获益。其核心亮点在于极高的功能丰富度：不仅支持文生图、图生图、局部重绘（Inpainting）和外绘（Outpainting）等基础模式，还独创了注意力机制调整、提示词矩阵、负向提示词以及“高清修复”等高级功能。此外，它内置了 GFPGAN 和 CodeFormer 等人脸修复工具，支持多种神经网络放大算法，并允许用户通过插件系统无限扩展能力。即使是显存有限的设备，stable-diffusion-webui 也提供了相应的优化选项，让高质量的 AI 艺术创作变得触手可及。",162132,"2026-04-05T11:01:52",[14,15,13],{"id":27,"name":28,"github_repo":29,"description_zh":30,"stars":31,"difficulty_score":32,"last_commit_at":33,"category_tags":34,"status":17},1381,"everything-claude-code","affaan-m\u002Feverything-claude-code","everything-claude-code 是一套专为 AI 编程助手（如 Claude Code、Codex、Cursor 等）打造的高性能优化系统。它不仅仅是一组配置文件，而是一个经过长期实战打磨的完整框架，旨在解决 AI 代理在实际开发中面临的效率低下、记忆丢失、安全隐患及缺乏持续学习能力等核心痛点。\n\n通过引入技能模块化、直觉增强、记忆持久化机制以及内置的安全扫描功能，everything-claude-code 能显著提升 AI 在复杂任务中的表现，帮助开发者构建更稳定、更智能的生产级 AI 代理。其独特的“研究优先”开发理念和针对 Token 消耗的优化策略，使得模型响应更快、成本更低，同时有效防御潜在的攻击向量。\n\n这套工具特别适合软件开发者、AI 研究人员以及希望深度定制 AI 工作流的技术团队使用。无论您是在构建大型代码库，还是需要 AI 协助进行安全审计与自动化测试，everything-claude-code 都能提供强大的底层支持。作为一个曾荣获 Anthropic 黑客大奖的开源项目，它融合了多语言支持与丰富的实战钩子（hooks），让 AI 真正成长为懂上",158594,2,"2026-04-16T23:34:05",[14,13,35],"语言模型",{"id":37,"name":38,"github_repo":39,"description_zh":40,"stars":41,"difficulty_score":42,"last_commit_at":43,"category_tags":44,"status":17},8272,"opencode","anomalyco\u002Fopencode","OpenCode 是一款开源的 AI 编程助手（Coding Agent），旨在像一位智能搭档一样融入您的开发流程。它不仅仅是一个代码补全插件，而是一个能够理解项目上下文、自主规划任务并执行复杂编码操作的智能体。无论是生成全新功能、重构现有代码，还是排查难以定位的 Bug，OpenCode 都能通过自然语言交互高效完成，显著减少开发者在重复性劳动和上下文切换上的时间消耗。\n\n这款工具专为软件开发者、工程师及技术研究人员设计，特别适合希望利用大模型能力来提升编码效率、加速原型开发或处理遗留代码维护的专业人群。其核心亮点在于完全开源的架构，这意味着用户可以审查代码逻辑、自定义行为策略，甚至私有化部署以保障数据安全，彻底打破了传统闭源 AI 助手的“黑盒”限制。\n\n在技术体验上，OpenCode 提供了灵活的终端界面（Terminal UI）和正在测试中的桌面应用程序，支持 macOS、Windows 及 Linux 全平台。它兼容多种包管理工具，安装便捷，并能无缝集成到现有的开发环境中。无论您是追求极致控制权的资深极客，还是渴望提升产出的独立开发者，OpenCode 都提供了一个透明、可信",144296,1,"2026-04-16T14:50:03",[13,45],"插件",{"id":47,"name":48,"github_repo":49,"description_zh":50,"stars":51,"difficulty_score":32,"last_commit_at":52,"category_tags":53,"status":17},2271,"ComfyUI","Comfy-Org\u002FComfyUI","ComfyUI 是一款功能强大且高度模块化的视觉 AI 引擎，专为设计和执行复杂的 Stable Diffusion 图像生成流程而打造。它摒弃了传统的代码编写模式，采用直观的节点式流程图界面，让用户通过连接不同的功能模块即可构建个性化的生成管线。\n\n这一设计巧妙解决了高级 AI 绘图工作流配置复杂、灵活性不足的痛点。用户无需具备编程背景，也能自由组合模型、调整参数并实时预览效果，轻松实现从基础文生图到多步骤高清修复等各类复杂任务。ComfyUI 拥有极佳的兼容性，不仅支持 Windows、macOS 和 Linux 全平台，还广泛适配 NVIDIA、AMD、Intel 及苹果 Silicon 等多种硬件架构，并率先支持 SDXL、Flux、SD3 等前沿模型。\n\n无论是希望深入探索算法潜力的研究人员和开发者，还是追求极致创作自由度的设计师与资深 AI 绘画爱好者，ComfyUI 都能提供强大的支持。其独特的模块化架构允许社区不断扩展新功能，使其成为当前最灵活、生态最丰富的开源扩散模型工具之一，帮助用户将创意高效转化为现实。",108322,"2026-04-10T11:39:34",[14,15,13],{"id":55,"name":56,"github_repo":57,"description_zh":58,"stars":59,"difficulty_score":32,"last_commit_at":60,"category_tags":61,"status":17},6121,"gemini-cli","google-gemini\u002Fgemini-cli","gemini-cli 是一款由谷歌推出的开源 AI 命令行工具，它将强大的 Gemini 大模型能力直接集成到用户的终端环境中。对于习惯在命令行工作的开发者而言，它提供了一条从输入提示词到获取模型响应的最短路径，无需切换窗口即可享受智能辅助。\n\n这款工具主要解决了开发过程中频繁上下文切换的痛点，让用户能在熟悉的终端界面内直接完成代码理解、生成、调试以及自动化运维任务。无论是查询大型代码库、根据草图生成应用，还是执行复杂的 Git 操作，gemini-cli 都能通过自然语言指令高效处理。\n\n它特别适合广大软件工程师、DevOps 人员及技术研究人员使用。其核心亮点包括支持高达 100 万 token 的超长上下文窗口，具备出色的逻辑推理能力；内置 Google 搜索、文件操作及 Shell 命令执行等实用工具；更独特的是，它支持 MCP（模型上下文协议），允许用户灵活扩展自定义集成，连接如图像生成等外部能力。此外，个人谷歌账号即可享受免费的额度支持，且项目基于 Apache 2.0 协议完全开源，是提升终端工作效率的理想助手。",100752,"2026-04-10T01:20:03",[45,13,15,14],{"id":63,"github_repo":64,"name":65,"description_en":66,"description_zh":67,"ai_summary_zh":68,"readme_en":69,"readme_zh":70,"quickstart_zh":71,"use_case_zh":72,"hero_image_url":73,"owner_login":74,"owner_name":75,"owner_avatar_url":76,"owner_bio":77,"owner_company":78,"owner_location":79,"owner_email":80,"owner_twitter":81,"owner_website":82,"owner_url":83,"languages":80,"stars":84,"forks":85,"last_commit_at":86,"license":87,"difficulty_score":42,"env_os":88,"env_gpu":89,"env_ram":89,"env_deps":90,"category_tags":93,"github_topics":94,"view_count":32,"oss_zip_url":80,"oss_zip_packed_at":80,"status":17,"created_at":98,"updated_at":99,"faqs":100,"releases":101},8245,"pat-jj\u002FAwesome-Adaptation-of-Agentic-AI","Awesome-Adaptation-of-Agentic-AI","Repo for \"Adaptation of Agentic AI\"","Awesome-Adaptation-of-Agentic-AI 是一个专注于“智能体自适应”领域的精选论文与资源库。随着 AI 智能体在复杂任务中广泛应用，如何让它们灵活适应新工具、新环境成为关键挑战。该项目系统梳理了智能体自适应的核心策略，将其分为“智能体自适应”与“工具自适应”两大方向，并进一步细分为基于工具执行信号、输出信号、代理无关及代理监督等具体技术路径。\n\n它解决了当前研究分散、缺乏统一分类框架的问题，帮助从业者快速定位前沿方法，如基于强化学习（GRPO、AlphaZero 变体）的工具调用优化、形式化证明中的编译器适配、以及多模态智能体在网页生成等场景中的动态调整机制。资源库不仅收录最新 arXiv 预印本和顶会论文，还附带代码链接与发展时间线，便于复现与对比。\n\n特别适合 AI 研究人员、大模型开发者及智能体系统架构师使用。无论是希望深入理解智能体如何通过反馈信号自我进化，还是寻找可落地的工具适配方案，都能从中获得清晰的技术地图。项目由多篇前沿工作支撑，持续更新，欢迎社区贡献。通过结构化整理与可视化呈现，Awesome-Adaptation-of-Agentic-AI","Awesome-Adaptation-of-Agentic-AI 是一个专注于“智能体自适应”领域的精选论文与资源库。随着 AI 智能体在复杂任务中广泛应用，如何让它们灵活适应新工具、新环境成为关键挑战。该项目系统梳理了智能体自适应的核心策略，将其分为“智能体自适应”与“工具自适应”两大方向，并进一步细分为基于工具执行信号、输出信号、代理无关及代理监督等具体技术路径。\n\n它解决了当前研究分散、缺乏统一分类框架的问题，帮助从业者快速定位前沿方法，如基于强化学习（GRPO、AlphaZero 变体）的工具调用优化、形式化证明中的编译器适配、以及多模态智能体在网页生成等场景中的动态调整机制。资源库不仅收录最新 arXiv 预印本和顶会论文，还附带代码链接与发展时间线，便于复现与对比。\n\n特别适合 AI 研究人员、大模型开发者及智能体系统架构师使用。无论是希望深入理解智能体如何通过反馈信号自我进化，还是寻找可落地的工具适配方案，都能从中获得清晰的技术地图。项目由多篇前沿工作支撑，持续更新，欢迎社区贡献。通过结构化整理与可视化呈现，Awesome-Adaptation-of-Agentic-AI 为构建更鲁棒、灵活的下一代智能体系统提供了坚实的知识基础。","# Awesome Adaptation of Agentic AI\n[![Awesome](https:\u002F\u002Fcdn.rawgit.com\u002Fsindresorhus\u002Fawesome\u002Fd7305f38d29fed78fa85652e3a63e154dd8e8829\u002Fmedia\u002Fbadge.svg)](https:\u002F\u002Fgithub.com\u002Fsindresorhus\u002Fawesome)\n[![Stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fpat-jj\u002FAwesome-Adaptation-of-Agentic-AI?style=social)](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fpat-jj\u002FAwesome-Adaptation-of-Agentic-AI?style=social)\n[![License: CC BY-NC-ND 4.0](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FLicense-CC--BY--NC--ND%204.0-blue.svg)](https:\u002F\u002Fcreativecommons.org\u002Flicenses\u002Fby-nc-nd\u002F4.0\u002F)\n[![PRWelcome](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FPRs-Welcome-red)](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FPRs-Welcome-red)\n[![arXiv](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FarXiv-2512.16301-b31b1b.svg)](https:\u002F\u002Farxiv.org\u002Fabs\u002F2512.16301)\n\n\u003Cp align=\"center\">\n    \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_d03e0022c5f7.png\" width=\"100%\" style=\"align:center;\"\u002F>\n\u003C\u002Fp>\n\nA curated list of papers on adaptation strategies of agentic AI systems. This repository accompanies the paper \"Adaptation of Agentic AI\" (Ongoing Work).\n\n**Cite this paper:**\n```\n@article{jiang2025adaptation,\n  title={Adaptation of Agentic AI},\n  author={Jiang, Pengcheng and Lin, Jiacheng and Shi, Zhiyi and Wang, Zifeng and He, Luxi and Wu, Yichen and Zhong, Ming and Song, Peiyang and Zhang, Qizheng and Wang, Heng and others},\n  journal={arXiv preprint arXiv:2512.16301},\n  year={2025}\n}\n```\n\n\n\n\n## Table of Contents\n- [Agent Adaptation](#agent-adaptation)\n  - [A1: Tool Execution Signaled](#a1-tool-execution-signaled)\n  - [A2: Agent Output Signaled](#a2-agent-output-signaled)\n- [Tool Adaptation](#tool-adaptation)\n  - [T1: Agent-Agnostic Tool Adaptation](#t1-agent-agnostic-tool-adaptation)\n  - [T2: Agent-Supervised Tool Adaptation](#t2-agent-supervised-tool-adaptation)\n\n---\n\n## Agent Adaptation\n\n### A1: Tool Execution Signaled Agent Adaptation\n\n\n\u003Cp align=\"center\">\n    \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_90e6987cdc1d.png\" width=\"100%\" style=\"align:center;\"\u002F>\n\u003C\u002Fp>\n\nDevelopment Timeline:\n\u003Cp align=\"center\">\n    \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_77f3db9358be.png\" width=\"100%\" style=\"align:center;\"\u002F>\n\u003C\u002Fp>\n\n#### RL-based Methods\n\n| Time | Method | Venue | Task(s) | Tool(s) | Agent Backbone | Tuning |\n|------|--------|-------|---------|---------|----------------|--------|\n| 2025.11 | Orion | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2511.07581) | IR | Retrievers | LFM2 | GRPO |\n| 2025.10 | olmOCR2 | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2510.19817)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fallenai\u002Folmocr) | Document OCR | Synthetic Document Verifier | Qwen2.5-VL | SFT, GRPO |\n| 2025.10 | AlphaProof | Nature’25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Fwww.nature.com\u002Farticles\u002Fs41586-025-09833-y) | Formal Theorem Proving | Lean Compiler | Transformer (3B Enc-Dec) | SFT, AlphaZero, TTRL |\n| 2025.10 | ToolExpander | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2510.07737) | Tool-Calling | Various APIs | Qwen2.5 | SFT, GRPO |\n| 2025.09 | BFS-Prover-V2 | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2509.06493)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002FByteDance-Seed\u002FBFS-Prover-V2) | Formal Theorem Proving | Lean Compiler | Qwen2.5 | BFS-Guided AlphaZero-like EI |\n| 2025.09 | WebGen-Agent | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2509.22644)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fmnluzimu\u002FWebGen-Agent) | Website Generation | VLM, GUI Agent, Code Executor | Various Models | SFT, Step-GRPO |\n| 2025.09 | Tool-R1 | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2509.12867)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002FYBYBZhang\u002FTool-R1) | General Tool-Augmented Reasoning, Multimodal QA | Code Execution, Multimedia Tools | Qwen2.5 | GRPO |\n| 2025.08 | FTRL | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2508.08791)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fbytedance\u002FFTRL) | Multi-Step Tool-Use | Simulated APIs | Qwen3 | GRPO |\n| 2025.08 | Goedel-Prover-V2 | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2508.03613)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002FGoedel-LM\u002FGoedel-Prover-V2) | Formal Theorem Proving | Lean Compiler | Qwen3 | SFT, GRPO |\n| 2025.07 | Leanabell-Prover-V2 | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2507.08649)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002FLeanabell-LM\u002FLeanabell-Prover-V2) | Formal Theorem Proving | Lean Compiler | Qwen2.5 | SFT, AlphaZero-like EI |\n| 2025.06 | Router-R1 | NeurIPS'25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2506.09033)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fulab-uiuc\u002FRouter-R1) | Multi-Round Routing | LLM Routing Pool | Qwen2.5, LLaMA3.2 | PPO |\n| 2025.05 | R1-Code-Interpreter | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2505.21668)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fyongchao98\u002FR1-Code-Interpreter) | Coding | Code Execution Sandbox | Qwen2.5 | GRPO |\n| 2025.05 | Tool-N1 | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2505.00024)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002FNVlabs\u002FTool-N1) | Tool-Calling | Various APIs | Qwen2.5 | GRPO |\n| 2025.04 | DeepSeek-Prover-V2 | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2504.21801)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fdeepseek-ai\u002FDeepSeek-Prover-V2) | Formal Theorem Proving | Lean Compiler | DeepSeek-V2 | SFT, GRPO |\n| 2025.04 | Kimina-Prover | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2504.11354)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fproject-numina\u002Fkimina-prover-rl) | Formal Theorem Proving | Lean Compiler | LLaMA-2 | SFT, AlphaZero-like EI |\n| 2025.04 | SQL-R1 | NeurIPS'25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2504.08600)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002FDataArcTech\u002FSQL-R1) | Text2SQL Search | SQL Engine | Qwen2.5, OmniSQL | SFT, GRPO |\n| 2025.03 | Rec-R1 | TMLR'25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Fopenreview.net\u002Fforum?id=YBRU9MV2vE)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Flinjc16\u002FRec-R1) | Recommendation Optimization | Recommendation System | Qwen2.5, LLaMA3.2 | GRPO |\n| 2025.03 | ReZero | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2504.11001)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fjanhq\u002FReZero) | Web Search, IR | Web Search Engine | LLaMA3.2 | GRPO |\n| 2025.03 | Code-R1 | ---\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fganler\u002Fcode-r1) | Coding | Code Executor | Qwen2.5 | GRPO |\n| 2025.02 | DeepRetrieval | COLM'25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2503.00223)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fpat-jj\u002FDeepRetrieval) | Web Search, IR, Text2SQL | Search Engine, Retrievers, SQL exec. | Qwen2.5, LLaMA3.2 | PPO, GRPO |\n| 2025.01 | DeepSeek-R1-Zero (Code) | Nature\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2501.12948) | Coding | Code Executor | DeepSeek-V3-Base | GRPO |\n| 2024.10 | RLEF | ICML'25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2410.02089) | Coding | Code Executor | LLaMA3.1 | PPO |\n| 2024.08 | DeepSeek-Prover-V1.5 | ICLR’25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2408.08152)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fdeepseek-ai\u002FDeepSeek-Prover-V1.5) | Formal Theorem Proving | Lean 4 Prover | DeepSeek-Prover-V1.5-RL | SFT, GRPO |\n| 2024.05 | LeDex | NeurIPS'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2405.18649) | Coding | Code Executor | StarCoder & CodeLlaMA | SFT, PPO |\n\n\n#### SFT & DPO Methods\n\n| Time | Method | Venue | Task(s) | Tool(s) | Agent Backbone | Tuning |\n|------|--------|-------|---------|---------|----------------|--------|\n| 2024.12 | AWL | ICML'25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2411.00412)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code]([YOUR_GITHUB_LINK](https:\u002F\u002Fgithub.com\u002FRose-STL-Lab\u002FAdapting-While-Learning)) | Scientific Reasoning,\u003Cbr>Adaptive Tool Usage | Scientific Simulators | Llama-3.1-8B,\u003Cbr>Qwen-2.5-{14\u002F32}B | SFT, DPO |\n| 2024.10 | LeReT | ICLR'25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2410.23214)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fsher222\u002FLeReT) | IR | Dense Retriever | LLaMA3, Gemma2 | DPO-like (IPO) |\n| 2024.10 | ToolFlow | NAACL'25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2410.18447) | Tool-Calling | Various APIs | LLaMA3.1 | SFT |\n| 2024.06 | TP-LLaMA | NeurIPS'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2406.07115) | Tool-Calling | Various APIs | LLaMA2 | SFT, DPO |\n| 2024.05 | AutoTools | WWW'25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2405.16533)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fmangopy\u002FAutoTools) | Automated Tool-Calling | Various APIs | GPT4, LLaMA3, Mistral | SFT |\n| 2024.03 | CYCLE | OOPSLA'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2403.18746) | Coding | Code Executor | CodeGen, StarCoder | SFT |\n| 2024.02 | RetPO | NAACL'25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2402.11827)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fdmis-lab\u002FRetPO) | IR | Retriever | LLaMA2-7B | SFT, DPO |\n| 2024.02 | CodeAct | ICML'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2402.01030)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fxingyaoww\u002Fcode-act) | Coding | Code Executor | LLaMA2, Mistral | SFT |\n| 2024.01 | NExT | ICML'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2404.14662) | Program Repair | Code Executor | PaLM2 | SFT |\n| 2023.07 | ToolLLM | ICLR'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2307.16789)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002FOpenBMB\u002FToolBench) | Tool-Calling, API Planning, Multi-Tool Reasoning | Real-World APIs | LLaMA, Vicuna | SFT |\n| 2023.06 | ToolAlpaca | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2306.05301)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Ftangqiaoyu\u002FToolAlpaca) | Multi-Turn Tool-Use | Simulated APIs | Vicuna | SFT |\n| 2023.05 | Gorilla | NeurIPS'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2305.15334)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002FShishirPatil\u002Fgorilla) | Tool-Calling, API Retrieval | Various APIs | LLaMA | SFT |\n| 2023.05 | TRICE | NAACL'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2305.13068)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fzjunlp\u002FTRICE) | Math Reasoning, QA, Multilingual QA, Knowledge Retrieval | Calculator, WikiSearch, Atlas QA Model, NLLB Translator | ChatGLM, Alpaca, Vicuna | SFT |\n| 2023.02 | Toolformer | NeurIPS'23\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2302.04761)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fconceptofmind\u002Ftoolformer) | QA, Math | Calculator, QA system, Search Engine, Translation System, Calendar | GPT-J | SFT |\n\n---\n\n\u003Cbr>\n\u003Cbr>\n\n### A2: Agent Output Signaled Agent Adaptation\n\n\u003Cp align=\"center\">\n    \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_9803f977b6c8.png\" width=\"100%\" style=\"align:center;\"\u002F>\n\u003C\u002Fp>\n\nDevelopment Timeline:\n\u003Cp align=\"center\">\n    \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_c786518ae572.png\" width=\"100%\" style=\"align:center;\"\u002F>\n\u003C\u002Fp>\n\n\n#### Methods with Tools\n\n| Time | Method | Venue | Task(s) | Tool(s) | Agent Backbone | Tuning |\n|------|--------|-------|---------|---------|----------------|--------|\n| 2025.10 | TT-SI | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2510.07841) | Tool Calling | Various APIs | Qwen2.5 | Test-Time Fine-Tuning |\n| 2025.10 | A²FM | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2510.12838)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002FOPPO-PersonalAI\u002FAdaptive_Agent_Foundation_Models) | Web Navigation, Math, QA | Search Engine, Crawl, Code Executor | Qwen2.5 | APO, GRPO |\n| 2025.09 | VerlTool | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2509.01055)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002FTIGER-AI-Lab\u002Fverl-tool) | Math, QA, SQL, Visual, Web Search, Coding | Code Interpreter, Search Engine, SQL Executor, Vision Tools | Qwen2.5, Qwen3 | GRPO |\n| 2025.08 | MedResearcher-R1 | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2508.14880)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002FAQ-MedAI\u002FMedResearcher-R1) | Medical Multi-hop QA | Medical Retriever, Web Search API, Document Reader | MedResearcher-R1 | SFT, GRPO |\n| 2025.08 | Agent Lightning | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2508.03680)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fagent-lightning) | Text-to-SQL, RAG, Math | SQL Executor, Retriever, Calculator | LLaMA3.2 | LightningRL |\n| 2025.07 | CodePRM | ACL'25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Faclanthology.org\u002F2025.findings-acl.428\u002F) | Coding | Code Executor | Qwen2.5-Coder | SFT |\n| 2025.07 | DynaSearcher | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2507.17365)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fmodelscope.cn\u002Fhome) | Multi-Hop QA, RAG | Document Search, KG Search | Qwen2.5, LLaMA3.1 | GRPO |\n| 2025.06 | MMSearch-R1 | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2506.20670)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002FEvolvingLMMs-Lab\u002Fmultimodal-search-r1) | Web Browsing, QA, Multimodal Search | Image Search, Web Browsing, Retriever | Qwen2.5 | REINFORCE, SFT |\n| 2025.06 | Self-Challenging | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2506.01716) | Web Browsing, Calculation, Retail, Airline | Code Interpreter, Web Browser, Database APIs | LLaMA3.1 | REINFORCE, SFT |\n| 2025.05 | StepSearch | EMNLP'25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2505.15107)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002FZillwang\u002FStepSearch) | Multi-Hop QA | Search Engine, Retriever | Qwen2.5 | StePPO |\n| 2025.05 | ZeroSearch | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2505.04588)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002FAlibaba-NLP\u002FZeroSearch) | Multi-Hop QA, QA | Search Engine, Web Search | Qwen2.5, LLaMA3.2 | REINFORCE, GPRO, PPO, SFT |\n| 2025.05 | AutoRefine | NeurIPS'25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2505.11277)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fsyr-cn\u002FAutoRefine) | Multi-Hop QA, QA | Retriever | Qwen2.5 | GRPO |\n| 2025.04 | ReTool | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2504.11536)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002FReTool-RL\u002FReTool) | Math | Code Interpreter | Qwen2.5 | PPO |\n| 2025.04 | ToolRL | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2504.13958)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fqiancheng0\u002FToolRL) | Tool Calling | Various Tools | Various Models | GRPO |\n| 2025.04 | DeepResearcher | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2504.03160)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002FGAIR-NLP\u002FDeepResearcher) | QA, Multi-Hop Reasoning, Deep Research | Web Search API, Web Browser | Qwen2.5 | GRPO |\n| 2025.03 | ReSearch | NeurIPS'25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2503.19470)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002FAgent-RL\u002FReCall) | QA | Search Engine, Retriever | Qwen2.5 | GRPO |\n| 2025.03 | Search-R1 | COLM'25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2503.09516)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002FPeterGriffinJin\u002FSearch-R1) | QA | Search Engine, Retriever | Qwen2.5 | PPO, GRPO |\n| 2025.03 | R1-Searcher | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2503.05592)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002FRUCAIBox\u002FR1-Searcher) | QA | Retriever | LLaMA3.1, Qwen2.5 | REINFORCE++ |\n| 2025.02 | RAS | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2502.10996)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fpat-jj\u002FRAS) | QA | Retriever | LLaMA2, LLaMA3.2 | SFT |\n| 2025.01 | Agent-R | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2501.11425)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002FByteDance-Seed\u002FAgent-R) | Various Tasks | Monte Carlo Tree Search | Qwen2.5, LLaMA3.2 | SFT |\n| 2024.06 | Re-ReST | EMNLP'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2406.01495)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002FPlusLabNLP\u002FRe-ReST) | Multi-Hop QA, VQA, Sequential Decision, Coding | Various APIs | Various Models | DPO |\n| 2024.06 | RPG | EMNLP'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2406.14979)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fharuhi-sudo\u002FRPG) | RAG, QA, Multi-hop Reasoning | Search Engine, Retriever | LLaMA2, GPT3.5 | SFT |\n| 2023.10 | Self-RAG | ICLR'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2310.11511)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002FAkariAsai\u002Fself-rag) | RAG, QA, Fact Verification | Retriever | LLaMA2 | SFT |\n| 2023.10 | FireAct | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2310.05915)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Ffireact-agent.github.io) | QA | Search API | GPT3.5, LLaMA2, CodeLLaMA | SFT |\n\n#### Methods without Tools\n\n| Time | Method | Venue | Task(s) | Tool(s) | Agent Backbone | Tuning |\n|------|--------|-------|---------|---------|----------------|--------|\n| 2025.10 | Empower | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2510.13709)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Ffestusev\u002Fcodegen_empowerment\u002Ftree\u002Fmain) | Coding | --- | Gemma3 | SFT |\n| 2025.10 | KnowRL | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2510.11407)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fanonymous.4open.science\u002Fr\u002FKnowRL-5BF0) | Knowledge calibration | --- | LLaMA3.1, Qwen2.5 | REINFORCE++ |\n| 2025.10 | GRACE | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2510.04506)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002FGasolSun36\u002FGRACE) | Embedding Tasks | --- | Qwen2.5, Qwen3, LLaMA3.2 | GRPO |\n| 2025.06 | Magistral | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2506.10910) | Math, Coding | --- | Magistral | PPO, GRPO |\n| 2025.05 | EHRMind | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2505.24105)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Flinjc16\u002FEHRMind) | EHR-based Reasoning | --- | LLaMA3 | SFT, GRPO |\n| 2025.01 | Kimi k1.5 | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2501.12948)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002FMoonshotAI\u002FKimi-k1.5) | Math, Coding | --- | Kimi k1.5 | GRPO |\n| 2025.01 | DeepSeek-R1-Zero (Math) | Nature\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2501.12948) | Math | --- | DeepSeek-V3 | GRPO |\n| 2024.09 | SCoRe | ICLR'25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2409.12917)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002FBY571\u002FSCoRe) | Math, Coding, QA | --- | Gemini1.0 Pro, Gemini1.5 Flash | REINFORCE |\n| 2024.07 | RISE | NeurIPS'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2407.18219)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fcmu-mind\u002FRISE) | Math | --- | LLaMA2, LLaMA3, Mistral | SFT |\n| 2024.06 | TextGrad | Nature\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2406.07496)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fzou-group\u002Ftextgrad) | Various Tasks | --- | GPT3.5, GPT4o | Prompt Tuning |\n| 2023.03 | Self-Refine | NeurIPS'23\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2303.17651)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fmadaan\u002Fself-refine) | Dialogue, Math, Coding | --- | GPT3.5, GPT4, CODEX | Test-Time Prompting |\n\n---\n\n## Tool Adaptation\n\n### T1: Agent-Agnostic Tool Adaptation\n\u003Cp align=\"center\">\n    \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_6ce24742a51c.png\" width=\"100%\" style=\"align:center;\"\u002F>\n\u003C\u002Fp>\n\n\n#### Foundational Systems and Architectures\n\n| Year.Month | Method Name | Venue | Paper Name |\n|:-----------:|:-----------:|:-----------|:-----------|\n| 2021.08 | Neural Operators | JMLR'23\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Fjmlr.org\u002Fpapers\u002Fv24\u002F21-1524.html) | Neural Operator: Learning Maps Between Function Spaces |\n| 2023.09 | HuggingGPT | NeurIPS'23\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2303.17580)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002FJARVIS) | HuggingGPT: Solving AI Tasks with ChatGPT and its Friends in Hugging Face |\n| 2023.08 | ViperGPT | ICCV'23\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2303.08128)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fcvlab-columbia\u002Fviper) | ViperGPT: Visual Inference via Python Execution for Reasoning |\n| 2025.07 | SciToolAgent | Nature Comp. Sci.'25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2507.20280) | SciToolAgent: A Knowledge-Graph-Driven Scientific Agent for Multitool Integration |\n#### Categories and Training Methods\n\n| Year.Month | Method Name | Venue | Paper Name |\n|:-----------:|:-----------:|:-----------|:-----------|\n| 2021.01 | CLIP | ICML'21\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2103.00020)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fopenai\u002FCLIP) | Learning Transferable Visual Models from Natural Language Supervision |\n| 2023.04 | SAM | ICCV'23\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2304.02643)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Ffacebookresearch\u002Fsegment-anything) | Segment Anything |\n| 2024.06 | SAM-CLIP | CVPR'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2310.15308) | SAM-CLIP: Merging Vision Foundation Models Towards Semantic and Spatial Understanding |\n| 2023.12 | Whisper | ICML'23\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2212.04356)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fopenai\u002Fwhisper) | Robust Speech Recognition via Large-Scale Weak Supervision |\n| 2024.10 | CodeAct | ICML'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2402.01030)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fxingyaoww\u002Fcode-act) | Executable Code Actions Elicit Better LLM Agents |\n| 2020.04 | DPR | EMNLP'20\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2004.04906)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Ffacebookresearch\u002FDPR) | Dense Passage Retrieval for Open-Domain Question Answering |\n| 2020.04 | ColBERT | SIGIR'20\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2004.12832)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fstanford-futuredata\u002FColBERT) | ColBERT: Efficient and Effective Passage Search via Contextualized Late Interaction over BERT |\n| 2021.12 | Contriever | TMLR'22\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2112.09118)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Ffacebookresearch\u002Fcontriever) | Unsupervised Dense Information Retrieval with Contrastive Learning |\n| 2022.12 | e5 | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2212.03533)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Funilm\u002Ftree\u002Fmaster\u002Fe5) | Text Embeddings by Weakly-Supervised Contrastive Pre-training |\n| 2021.07 | AlphaFold2 | Nature\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Fwww.nature.com\u002Farticles\u002Fs41586-021-03819-2)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fdeepmind\u002Falphafold) | Highly Accurate Protein Structure Prediction with AlphaFold |\n| 2023.03 | ESMFold | Science\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Fwww.science.org\u002Fdoi\u002F10.1126\u002Fscience.ade2574) | Evolutionary-Scale Prediction of Atomic-Level Protein Structure with a Language Model |\n---\n\n\u003Cbr>\n\u003Cbr>\n\n### T2: Agent-Supervised Tool Adaptation\n\n\u003Cp align=\"center\">\n    \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_aa235c831c30.png\" width=\"100%\" style=\"align:center;\"\u002F>\n\u003C\u002Fp>\n\nDevelopment Timeline:\n\u003Cp align=\"center\">\n    \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_a11f65d23972.png\" width=\"100%\" style=\"align:center;\"\u002F>\n\u003C\u002Fp>\n\n| Time | Method | Venue | Task(s) | Tool Backbone | Agent Backbone | Tuning |\n|------|--------|-------|---------|---------------|----------------|--------|\n| 2025.10 | QAgent | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2510.08383)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002FLivingFutureLab\u002FQAgent) | QA, RAG | Qwen2.5-3B | Qwen-7B | GRPO |\n| 2025.10 | AgentFlow | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2510.05592)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Flupantech\u002FAgentFlow) | Web Search, Planning, Reasoning, Math | Qwen2.5-7B | Qwen2.5-7B | Flow-GRPO |\n| 2025.10 | Advisor Models | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2510.02453)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Faz1326\u002Fadvisor-models) | Math, Reasoning | Qwen2.5-7B, Qwen3-8B | GPT-4o-Mini, GPT-5, Claude4-Sonnet, GPT-4.1-Mini | GRPO |\n| 2025.10 | AutoGraph-R1 | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2510.15339)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002FHKUST-KnowComp\u002FAutoGraph-R1) | KG Construction, RAG | KG Constructor (Qwen2.5-3B\u002F7B) | Frozen RAG Generator (Qwen2.5-7B) | GRPO |\n| 2025.10 | MAE | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2510.23595)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fulab-uiuc\u002FMulti-agent-Evolve) | Math, Coding, Commonsense Reasoning | Qwen2.5-3B | Qwen2.5-3B | REINFORCE++ |\n| 2025.09 | Mem-α | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2509.25911)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fwangyu-ustc\u002FMem-alpha) | Retrieval, Test-Time Learning, Long-Range Understanding | Qwen3-4B | Qwen3-4B, Qwen3-32B, GPT-4.1-Mini | GRPO |\n| 2025.08 | AI-SearchPlanner | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2508.20368) | Web QA | Qwen3-32b | Qwen2.5-7B | PPO |\n| 2025.08 | Memento | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2508.16153)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002FAgent-on-the-Fly\u002FMemento) | Long-Horizon Reasoning, Web Research, QA, Academic Reasoning | Q-function (two-layer MLPs) | GPT-4.1 | Soft Q-Learning |\n| 2025.08 | R-Zero | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2508.05004)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002FChengsong-Huang\u002FR-Zero) | Math, Reasoning | Qwen3-4B, Qwen3-8B, OctoThinker-3B, OctoThinker-8B | Qwen3-4B, Qwen3-8B, OctoThinker-3B, OctoThinker-8B | GRPO |\n| 2025.06 | Sysformer | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2506.15751) | QA, RAG | Small Transformer | LLaMA-2-7B, LLaMA-3.1-8B, Mistral-7B, Phi-3.5-mini, Zephyr-7B-beta | Supervised Learning |\n| 2025.05 | s3 | EMNLP'25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2505.14146)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fpat-jj\u002Fs3) | QA, RAG | Qwen2.5-7B | Qwen2.5-7B, Qwen2.5-14B, Claude-3-Haiku | PPO |\n| 2024.10 | Matryoshka Pilot | NeurIPS'25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2410.20749)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Flichangh20\u002FMatryoshka) | Math, Planning, Reasoning | LLaMA3-8B, Qwen2.5-7B | GPT-4o-Mini, GPT-3.5-Turbo | DPO, IDPO |\n| 2024.06 | CoBB | EMNLP'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2406.18695)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fbbuing9\u002FCoBB) | QA, Math | Mistral-7b-inst-v2 | GPT-3.5-Turbo, Claude-3-Haiku, Phi-3-mini-4k-inst, Gemma-1.1-7B-it, Mistral-7B-inst-v2 | SFT, ORPO |\n| 2024.05 | Medadapter | EMNLP'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2405.03000)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fwshi83\u002FMedAdapter) | Medical QA, NLI, RQE | BERT-Base-Uncased | GPT-3.5-Turbo | SFT, BPO |\n| 2024.03 | BLADE | AAAI'25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2403.18365)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002FCSHaitao\u002FBLADE) | Domain-Specific QA | BLOOMZ-1b7 | ChatGPT, ChatGLM, Baichuan, Qwen | SFT, BPO |\n| 2024.02 | ARL2 | ACL'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2402.13542)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fzhanglingxi-cs\u002FARL2) | QA | LLaMA2-7B | GPT-3.5-Turbo | Contrastive Learning |\n| 2024.02 | EVOR | EMNLP'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2402.12317)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fxlang-ai\u002FEVOR) | RAG-based Coding | GPT-3.5-Turbo | GPT-3.5-Turbo, CodeLLaMA | Prompt Engineering |\n| 2024.02 | Bbox-Adapter | ICML'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2402.08219)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fhaotiansun14\u002FBBox-Adapter) | QA | DeBERTa-v3-base (0.1B), DeBERTa-v3-large (0.3B) | GPT-3.5-Turbo, Mixtral-8x7B | Contrastive Learning |\n| 2024.01 | Proxy-Tuning | COLM'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2401.08565)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Falisawuffles\u002Fproxy-tuning) | QA, Math, Code | LLaMA2-7B | LLaMA2-70B | Proxy-Tuning |\n| 2024.01 | BGM | ACL'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2401.06954) | QA, Personalized Generation (NQ, HotpotQA, Email, Book) | T5-XXL-11B | PaLM2-S | SFT, PPO |\n| 2023.10 | RA-DIT | ICLR'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2310.01352) | Knowledge-Intensive Tasks (MMLU, NQ, TQA, ELI5, HotpotQA, etc.) | DRAGON+ | LLaMA-65B | SFT, LSR |\n| 2023.06 | LLM-R | EACL'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2307.07164)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002FLMOps\u002Ftree\u002Fmain\u002Fllm_retriever) | Zero-shot NLU (Reading Comprehension, QA, NLI, Paraphrase, Sentiment, Summarization) | E5-base | GPT-Neo-2.7B, LLaMA-13B, GPT-3.5-Turbo | Contrastive Learning |\n| 2023.05 | AAR | ACL'23\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2305.17331)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002FOpenMatch\u002FAugmentation-Adapted-Retriever) | Zero-Shot Generalization (MMLU, PopQA) | ANCE, Contriever | Flan-T5-Small, InstructGPT | Contrastive Learning |\n| 2023.05 | ToolkenGPT | NeurIPS'23\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2305.11554)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002FBer666\u002FToolkenGPT) | Numerical Reasoning, QA, Plan Generation | Token Embedding | GPT-J 6B, OPT-6.7B, OPT-13B | Proxy-Tuning |\n| 2023.03 | UPRISE | EMNLP'23\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Faclanthology.org\u002F2023.emnlp-main.758\u002F)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002FLMOps) | Zero-shot NLU (Reading Comprehension, QA, NLI, Paraphrase, Sentiment, Summarization) | GPT-Neo-2.7B | BLOOM-7.1B, OPT-66B, GPT-3-175B | Contrastive Learning |\n| 2023.01 | REPLUG | NAACL'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [Paper](https:\u002F\u002Faclanthology.org\u002F2024.naacl-long.463.pdf)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [Code](https:\u002F\u002Fgithub.com\u002Fswj0419\u002FREPLUG) | QA | Contriever | GPT3-175B, PaLM, Codex, LLaMA-13B | Proxy-Tuning, LSR |\n\n---\n\n## Citation\n\nIf you find this repository useful, please consider citing our survey:\n\n```\n@article{jiang2025adaptation,\n  title={Adaptation of Agentic AI},\n  author={Jiang, Pengcheng and Lin, Jiacheng and Shi, Zhiyi and Wang, Zifeng and He, Luxi and Wu, Yichen and Zhong, Ming and Song, Peiyang and Zhang, Qizheng and Wang, Heng and others},\n  journal={arXiv preprint arXiv:2512.16301},\n  year={2025}\n}\n```\n\n## Contributing\n\nWe welcome contributions! Please feel free to submit a Pull Request to add new papers or update existing entries.\n\n---\n\n\u003Cp align=\"center\">\n  \u003Cbr>\n  \u003Csub>(ﾉ◕ヮ◕)ﾉ*:･ﾟ✧ Keep exploring the awesome world of agentic AI! ✧ﾟ･: *ヽ(◕ヮ◕ヽ)\u003C\u002Fsub>\n\u003C\u002Fp>\n","# 智能体AI的精彩适配\n[![Awesome](https:\u002F\u002Fcdn.rawgit.com\u002Fsindresorhus\u002Fawesome\u002Fd7305f38d29fed78fa85652e3a63e154dd8e8829\u002Fmedia\u002Fbadge.svg)](https:\u002F\u002Fgithub.com\u002Fsindresorhus\u002Fawesome)\n[![星标](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fpat-jj\u002FAwesome-Adaptation-of-Agentic-AI?style=social)](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fpat-jj\u002FAwesome-Adaptation-of-Agentic-AI?style=social)\n[![许可证：CC BY-NC-ND 4.0](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FLicense-CC--BY--NC--ND%204.0-blue.svg)](https:\u002F\u002Fcreativecommons.org\u002Flicenses\u002Fby-nc-nd\u002F4.0\u002F)\n[![欢迎PR](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FPRs-Welcome-red)](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FPRs-Welcome-red)\n[![arXiv](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FarXiv-2512.16301-b31b1b.svg)](https:\u002F\u002Farxiv.org\u002Fabs\u002F2512.16301)\n\n\u003Cp align=\"center\">\n    \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_d03e0022c5f7.png\" width=\"100%\" style=\"align:center;\"\u002F>\n\u003C\u002Fp>\n\n一份精心整理的关于智能体AI系统适应策略的论文列表。本仓库与论文《智能体AI的适应》（正在进行中）配套使用。\n\n**引用此论文：**\n```\n@article{jiang2025adaptation,\n  title={智能体AI的适应},\n  author={Jiang, Pengcheng and Lin, Jiacheng and Shi, Zhiyi and Wang, Zifeng and He, Luxi and Wu, Yichen and Zhong, Ming and Song, Peiyang and Zhang, Qizheng and Wang, Heng and others},\n  journal={arXiv预印本 arXiv:2512.16301},\n  year={2025}\n}\n```\n\n\n\n\n## 目录\n- [智能体适应](#agent-adaptation)\n  - [A1：工具执行信号驱动](#a1-tool-execution-signaled)\n  - [A2：智能体输出信号驱动](#a2-agent-output-signaled)\n- [工具适应](#tool-adaptation)\n  - [T1：与智能体无关的工具适应](#t1-agent-agnostic-tool-adaptation)\n  - [T2：由智能体监督的工具适应](#t2-agent-supervised-tool-adaptation)\n\n---\n\n## 智能体适应\n\n### A1：工具执行信号驱动的智能体适应\n\n\n\u003Cp align=\"center\">\n    \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_90e6987cdc1d.png\" width=\"100%\" style=\"align:center;\"\u002F>\n\u003C\u002Fp>\n\n开发时间线：\n\u003Cp align=\"center\">\n    \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_77f3db9358be.png\" width=\"100%\" style=\"align:center;\"\u002F>\n\u003C\u002Fp>\n\n#### 基于强化学习的方法\n\n| 时间 | 方法 | 场景 | 任务 | 工具 | 模型基座 | 微调方式 |\n|------|--------|-------|---------|---------|----------------|--------|\n| 2025.11 | Orion | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2511.07581) | 信息检索 | 检索器 | LFM2 | GRPO |\n| 2025.10 | olmOCR2 | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2510.19817)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fallenai\u002Folmocr) | 文档OCR | 合成文档验证器 | Qwen2.5-VL | SFT, GRPO |\n| 2025.10 | AlphaProof | Nature’25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Fwww.nature.com\u002Farticles\u002Fs41586-025-09833-y) | 形式化定理证明 | Lean编译器 | Transformer (3B 编解码) | SFT, AlphaZero, TTRL |\n| 2025.10 | ToolExpander | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2510.07737) | 工具调用 | 各类API | Qwen2.5 | SFT, GRPO |\n| 2025.09 | BFS-Prover-V2 | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2509.06493)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002FByteDance-Seed\u002FBFS-Prover-V2) | 形式化定理证明 | Lean编译器 | Qwen2.5 | BFS引导的类似AlphaZero的强化学习 |\n| 2025.09 | WebGen-Agent | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2509.22644)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fmnluzimu\u002FWebGen-Agent) | 网站生成 | VLM、GUI代理、代码执行器 | 多种模型 | SFT, Step-GRPO |\n| 2025.09 | Tool-R1 | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2509.12867)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002FYBYBZhang\u002FTool-R1) | 通用工具增强推理、多模态问答 | 代码执行、多媒体工具 | Qwen2.5 | GRPO |\n| 2025.08 | FTRL | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2508.08791)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fbytedance\u002FFTRL) | 多步工具使用 | 模拟API | Qwen3 | GRPO |\n| 2025.08 | Goedel-Prover-V2 | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2508.03613)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002FGoedel-LM\u002FGoedel-Prover-V2) | 形式化定理证明 | Lean编译器 | Qwen3 | SFT, GRPO |\n| 2025.07 | Leanabell-Prover-V2 | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2507.08649)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002FLeanabell-LM\u002FLeanabell-Prover-V2) | 形式化定理证明 | Lean编译器 | Qwen2.5 | SFT, 类似AlphaZero的强化学习 |\n| 2025.06 | Router-R1 | NeurIPS'25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2506.09033)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fulab-uiuc\u002FRouter-R1) | 多轮路由 | LLM路由池 | Qwen2.5、LLaMA3.2 | PPO |\n| 2025.05 | R1-Code-Interpreter | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2505.21668)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fyongchao98\u002FR1-Code-Interpreter) | 编程 | 代码执行沙箱 | Qwen2.5 | GRPO |\n| 2025.05 | Tool-N1 | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2505.00024)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002FNVlabs\u002FTool-N1) | 工具调用 | 各类API | Qwen2.5 | GRPO |\n| 2025.04 | DeepSeek-Prover-V2 | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2504.21801)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fdeepseek-ai\u002FDeepSeek-Prover-V2) | 形式化定理证明 | Lean编译器 | DeepSeek-V2 | SFT, GRPO |\n| 2025.04 | Kimina-Prover | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2504.11354)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fproject-numina\u002Fkimina-prover-rl) | 形式化定理证明 | Lean编译器 | LLaMA-2 | SFT, 类似AlphaZero的强化学习 |\n| 2025.04 | SQL-R1 | NeurIPS'25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2504.08600)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002FDataArcTech\u002FSQL-R1) | 文本转SQL查询 | SQL引擎 | Qwen2.5、OmniSQL | SFT, GRPO |\n| 2025.03 | Rec-R1 | TMLR'25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Fopenreview.net\u002Fforum?id=YBRU9MV2vE)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Flinjc16\u002FRec-R1) | 推荐优化 | 推荐系统 | Qwen2.5、LLaMA3.2 | GRPO |\n| 2025.03 | ReZero | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2504.11001)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fjanhq\u002FReZero) | 网络搜索、信息检索 | 网络搜索引擎 | LLaMA3.2 | GRPO |\n| 2025.03 | Code-R1 | ---\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fganler\u002Fcode-r1) | 编程 | 代码执行器 | Qwen2.5 | GRPO |\n| 2025.02 | DeepRetrieval | COLM'25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2503.00223)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fpat-jj\u002FDeepRetrieval) | 网络搜索、信息检索、文本转SQL | 搜索引擎、检索器、SQL执行器 | Qwen2.5、LLaMA3.2 | PPO、GRPO |\n| 2025.01 | DeepSeek-R1-Zero（代码） | Nature\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2501.12948) | 编程 | 代码执行器 | DeepSeek-V3-Base | GRPO |\n| 2024.10 | RLEF | ICML'25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2410.02089) | 编程 | 代码执行器 | LLaMA3.1 | PPO |\n| 2024.08 | DeepSeek-Prover-V1.5 | ICLR'25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2408.08152)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fdeepseek-ai\u002FDeepSeek-Prover-V1.5) | 形式化定理证明 | Lean 4 Prover | DeepSeek-Prover-V1.5-RL | SFT, GRPO |\n| 2024.05 | LeDex | NeurIPS'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2405.18649) | 编程 | 代码执行器 | StarCoder & CodeLlaMA | SFT, PPO |\n\n\n#### SFT与DPO方法\n\n| 时间 | 方法 | 会议\u002F期刊 | 任务 | 工具 | 模型基座 | 微调方式 |\n|------|--------|-------|---------|---------|----------------|--------|\n| 2024.12 | AWL | ICML'25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2411.00412)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码]([YOUR_GITHUB_LINK](https:\u002F\u002Fgithub.com\u002FRose-STL-Lab\u002FAdapting-While-Learning)) | 科学推理,\u003Cbr>自适应工具使用 | 科学模拟器 | Llama-3.1-8B,\u003Cbr>Qwen-2.5-{14\u002F32}B | SFT, DPO |\n| 2024.10 | LeReT | ICLR'25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2410.23214)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fsher222\u002FLeReT) | 信息检索 | 密集检索器 | LLaMA3, Gemma2 | 类DPO（IPO） |\n| 2024.10 | ToolFlow | NAACL'25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2410.18447) | 工具调用 | 各类API | LLaMA3.1 | SFT |\n| 2024.06 | TP-LLaMA | NeurIPS'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2406.07115) | 工具调用 | 各类API | LLaMA2 | SFT, DPO |\n| 2024.05 | AutoTools | WWW'25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2405.16533)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fmangopy\u002FAutoTools) | 自动化工具调用 | 各类API | GPT4, LLaMA3, Mistral | SFT |\n| 2024.03 | CYCLE | OOPSLA'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2403.18746) | 编程 | 代码执行器 | CodeGen, StarCoder | SFT |\n| 2024.02 | RetPO | NAACL'25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2402.11827)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fdmis-lab\u002FRetPO) | 信息检索 | 检索器 | LLaMA2-7B | SFT, DPO |\n| 2024.02 | CodeAct | ICML'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2402.01030)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fxingyaoww\u002Fcode-act) | 编程 | 代码执行器 | LLaMA2, Mistral | SFT |\n| 2024.01 | NExT | ICML'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2404.14662) | 程序修复 | 代码执行器 | PaLM2 | SFT |\n| 2023.07 | ToolLLM | ICLR'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2307.16789)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002FOpenBMB\u002FToolBench) | 工具调用、API规划、多工具推理 | 现实世界API | LLaMA, Vicuna | SFT |\n| 2023.06 | ToolAlpaca | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2306.05301)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Ftangqiaoyu\u002FToolAlpaca) | 多轮工具使用 | 模拟API | Vicuna | SFT |\n| 2023.05 | Gorilla | NeurIPS'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2305.15334)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002FShishirPatil\u002Fgorilla) | 工具调用、API检索 | 各类API | LLaMA | SFT |\n| 2023.05 | TRICE | NAACL'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2305.13068)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fzjunlp\u002FTRICE) | 数学推理、问答、多语言问答、知识检索 | 计算器、WikiSearch、Atlas QA模型、NLLB翻译器 | ChatGLM, Alpaca, Vicuna | SFT |\n| 2023.02 | Toolformer | NeurIPS'23\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2302.04761)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fconceptofmind\u002Ftoolformer) | 问答、数学 | 计算器、问答系统、搜索引擎、翻译系统、日历 | GPT-J | SFT |\n\n---\n\n\u003Cbr>\n\u003Cbr>\n\n\n\n### A2: 基于信号的智能体自适应\n\n\u003Cp align=\"center\">\n    \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_9803f977b6c8.png\" width=\"100%\" style=\"align:center;\"\u002F>\n\u003C\u002Fp>\n\n开发时间线：\n\u003Cp align=\"center\">\n    \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_c786518ae572.png\" width=\"100%\" style=\"align:center;\"\u002F>\n\u003C\u002Fp>\n\n\n#### 带工具的方法\n\n| 时间 | 方法 | 论文出处 | 任务 | 工具 | 模型基座 | 微调方式 |\n|------|--------|-------|---------|---------|----------------|--------|\n| 2025.10 | TT-SI | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2510.07841) | 工具调用 | 各类API | Qwen2.5 | 测试时微调 |\n| 2025.10 | A²FM | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2510.12838)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002FOPPO-PersonalAI\u002FAdaptive_Agent_Foundation_Models) | 网页导航、数学、问答 | 搜索引擎、爬虫、代码执行器 | Qwen2.5 | APO、GRPO |\n| 2025.09 | VerlTool | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2509.01055)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002FTIGER-AI-Lab\u002Fverl-tool) | 数学、问答、SQL、视觉、网页搜索、编程 | 代码解释器、搜索引擎、SQL执行器、视觉工具 | Qwen2.5、Qwen3 | GRPO |\n| 2025.08 | MedResearcher-R1 | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2508.14880)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002FAQ-MedAI\u002FMedResearcher-R1) | 医疗多跳问答 | 医疗检索器、网页搜索API、文档阅读器 | MedResearcher-R1 | SFT、GRPO |\n| 2025.08 | Agent Lightning | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2508.03680)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fagent-lightning) | 文本转SQL、RAG、数学 | SQL执行器、检索器、计算器 | LLaMA3.2 | LightningRL |\n| 2025.07 | CodePRM | ACL'25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Faclanthology.org\u002F2025.findings-acl.428\u002F) | 编程 | 代码执行器 | Qwen2.5-Coder | SFT |\n| 2025.07 | DynaSearcher | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2507.17365)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fmodelscope.cn\u002Fhome) | 多跳问答、RAG | 文档搜索、知识图谱搜索 | Qwen2.5、LLaMA3.1 | GRPO |\n| 2025.06 | MMSearch-R1 | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2506.20670)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002FEvolvingLMMs-Lab\u002Fmultimodal-search-r1) | 网页浏览、问答、多模态搜索 | 图像搜索、网页浏览、检索器 | Qwen2.5 | REINFORCE、SFT |\n| 2025.06 | Self-Challenging | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2506.01716) | 网页浏览、计算、零售、航空 | 代码解释器、网页浏览器、数据库API | LLaMA3.1 | REINFORCE、SFT |\n| 2025.05 | StepSearch | EMNLP'25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2505.15107)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002FZillwang\u002FStepSearch) | 多跳问答 | 搜索引擎、检索器 | Qwen2.5 | StePPO |\n| 2025.05 | ZeroSearch | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2505.04588)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002FAlibaba-NLP\u002FZeroSearch) | 多跳问答、问答 | 搜索引擎、网页搜索 | Qwen2.5、LLaMA3.2 | REINFORCE、GPRO、PPO、SFT |\n| 2025.05 | AutoRefine | NeurIPS'25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2505.11277)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fsyr-cn\u002FAutoRefine) | 多跳问答、问答 | 检索器 | Qwen2.5 | GRPO |\n| 2025.04 | ReTool | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2504.11536)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002FReTool-RL\u002FReTool) | 数学 | 代码解释器 | Qwen2.5 | PPO |\n| 2025.04 | ToolRL | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2504.13958)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fqiancheng0\u002FToolRL) | 工具调用 | 各类工具 | 多种模型 | GRPO |\n| 2025.04 | DeepResearcher | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2504.03160)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002FGAIR-NLP\u002FDeepResearcher) | 问答、多跳推理、深度研究 | 网页搜索API、网页浏览器 | Qwen2.5 | GRPO |\n| 2025.03 | ReSearch | NeurIPS'25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2503.19470)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002FAgent-RL\u002FReCall) | 问答 | 搜索引擎、检索器 | Qwen2.5 | GRPO |\n| 2025.03 | Search-R1 | COLM'25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2503.09516)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002FPeterGriffinJin\u002FSearch-R1) | 问答 | 搜索引擎、检索器 | Qwen2.5 | PPO、GRPO |\n| 2025.03 | R1-Searcher | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2503.05592)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002FRUCAIBox\u002FR1-Searcher) | 问答 | 检索器 | LLaMA3.1、Qwen2.5 | REINFORCE++ |\n| 2025.02 | RAS | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2502.10996)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fpat-jj\u002FRAS) | 问答 | 检索器 | LLaMA2、LLaMA3.2 | SFT |\n| 2025.01 | Agent-R | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2501.11425)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002FByteDance-Seed\u002FAgent-R) | 多种任务 | 蒙特卡洛树搜索 | Qwen2.5、LLaMA3.2 | SFT |\n| 2024.06 | Re-ReST | EMNLP'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2406.01495)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002FPlusLabNLP\u002FRe-ReST) | 多跳问答、VQA、序列决策、编程 | 各类API | 多种模型 | DPO |\n| 2024.06 | RPG | EMNLP'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2406.14979)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fharuhi-sudo\u002FRPG) | RAG、问答、多跳推理 | 搜索引擎、检索器 | LLaMA2、GPT3.5 | SFT |\n| 2023.10 | Self-RAG | ICLR'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2310.11511)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002FAkariAsai\u002Fself-rag) | RAG、问答、事实核查 | 检索器 | LLaMA2 | SFT |\n| 2023.10 | FireAct | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2310.05915)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Ffireact-agent.github.io) | 问答 | 搜索API | GPT3.5、LLaMA2、CodeLLaMA | SFT |\n\n#### 无工具方法\n\n| 时间 | 方法 | 论文\u002F代码来源 | 任务 | 工具 | 模型基座 | 微调方式 |\n|------|--------|-------|---------|---------|----------------|--------|\n| 2025.10 | Empower | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2510.13709)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Ffestusev\u002Fcodegen_empowerment\u002Ftree\u002Fmain) | 编码 | --- | Gemma3 | SFT |\n| 2025.10 | KnowRL | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2510.11407)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fanonymous.4open.science\u002Fr\u002FKnowRL-5BF0) | 知识校准 | --- | LLaMA3.1, Qwen2.5 | REINFORCE++ |\n| 2025.10 | GRACE | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2510.04506)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002FGasolSun36\u002FGRACE) | 嵌入任务 | --- | Qwen2.5, Qwen3, LLaMA3.2 | GRPO |\n| 2025.06 | Magistral | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2506.10910) | 数学、编码 | --- | Magistral | PPO、GRPO |\n| 2025.05 | EHRMind | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2505.24105)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Flinjc16\u002FEHRMind) | 基于电子健康记录的推理 | --- | LLaMA3 | SFT、GRPO |\n| 2025.01 | Kimi k1.5 | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2501.12948)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002FMoonshotAI\u002FKimi-k1.5) | 数学、编码 | --- | Kimi k1.5 | GRPO |\n| 2025.01 | DeepSeek-R1-Zero (数学) | Nature\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2501.12948) | 数学 | --- | DeepSeek-V3 | GRPO |\n| 2024.09 | SCoRe | ICLR'25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2409.12917)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002FBY571\u002FSCoRe) | 数学、编码、问答 | --- | Gemini1.0 Pro、Gemini1.5 Flash | REINFORCE |\n| 2024.07 | RISE | NeurIPS'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2407.18219)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fcmu-mind\u002FRISE) | 数学 | --- | LLaMA2、LLaMA3、Mistral | SFT |\n| 2024.06 | TextGrad | Nature\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2406.07496)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fzou-group\u002Ftextgrad) | 多种任务 | --- | GPT3.5、GPT4o | 提示词微调 |\n| 2023.03 | Self-Refine | NeurIPS'23\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2303.17651)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fmadaan\u002Fself-refine) | 对话、数学、编码 | --- | GPT3.5、GPT4、CODEX | 测试时提示工程 |\n\n---\n\n\n\n## 工具适配\n\n### T1：代理无关的工具适配\n\u003Cp align=\"center\">\n    \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_6ce24742a51c.png\" width=\"100%\" style=\"align:center;\"\u002F>\n\u003C\u002Fp>\n\n\n#### 基础系统与架构\n\n| 年份.月份 | 方法名称 | 会议\u002F期刊 | 论文名称 |\n|:-----------:|:-----------:|:-----------|:-----------|\n| 2021.08 | 神经算子 | JMLR'23\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Fjmlr.org\u002Fpapers\u002Fv24\u002F21-1524.html) | 神经算子：学习函数空间之间的映射 |\n| 2023.09 | HuggingGPT | NeurIPS'23\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2303.17580)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002FJARVIS) | HuggingGPT：利用ChatGPT及其在Hugging Face中的伙伴解决AI任务 |\n| 2023.08 | ViperGPT | ICCV'23\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2303.08128)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fcvlab-columbia\u002Fviper) | ViperGPT：通过Python执行进行视觉推理 |\n| 2025.07 | SciToolAgent | Nature Comp. Sci.'25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2507.20280) | SciToolAgent：一种基于知识图谱的科学智能体，用于多工具集成 |\n#### 分类与训练方法\n\n| 年份.月份 | 方法名称 | 会议\u002F期刊 | 论文名称 |\n|:-----------:|:-----------:|:-----------|:-----------|\n| 2021.01 | CLIP | ICML'21\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2103.00020)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fopenai\u002FCLIP) | 从自然语言监督中学习可迁移的视觉模型 |\n| 2023.04 | SAM | ICCV'23\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2304.02643)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Ffacebookresearch\u002Fsegment-anything) | 分割一切 |\n| 2024.06 | SAM-CLIP | CVPR'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2310.15308) | SAM-CLIP：融合视觉基础模型以实现语义与空间理解 |\n| 2023.12 | Whisper | ICML'23\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2212.04356)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fopenai\u002Fwhisper) | 通过大规模弱监督实现鲁棒语音识别 |\n| 2024.10 | CodeAct | ICML'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2402.01030)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fxingyaoww\u002Fcode-act) | 可执行代码动作能够引导出更好的LLM智能体 |\n| 2020.04 | DPR | EMNLP'20\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2004.04906)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Ffacebookresearch\u002FDPR) | 面向开放域问答的密集段落检索 |\n| 2020.04 | ColBERT | SIGIR'20\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2004.12832)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fstanford-futuredata\u002FColBERT) | ColBERT：基于BERT的上下文化晚期交互实现高效且有效的段落搜索 |\n| 2021.12 | Contriever | TMLR'22\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2112.09118)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Ffacebookresearch\u002Fcontriever) | 基于对比学习的无监督密集信息检索 |\n| 2022.12 | e5 | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2212.03533)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Funilm\u002Ftree\u002Fmaster\u002Fe5) | 通过弱监督对比预训练生成文本嵌入 |\n| 2021.07 | AlphaFold2 | Nature\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Fwww.nature.com\u002Farticles\u002Fs41586-021-03819-2)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fdeepmind\u002Falphafold) | AlphaFold：高精度蛋白质结构预测 |\n| 2023.03 | ESMFold | Science\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Fwww.science.org\u002Fdoi\u002F10.1126\u002Fscience.ade2574) | 利用语言模型进行原子级蛋白质结构的进化尺度预测 |\n---\n\n\u003Cbr>\n\u003Cbr>\n\n### T2：代理监督的工具适配\n\n\u003Cp align=\"center\">\n    \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_aa235c831c30.png\" width=\"100%\" style=\"align:center;\"\u002F>\n\u003C\u002Fp>\n\n开发时间线：\n\u003Cp align=\"center\">\n    \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_a11f65d23972.png\" width=\"100%\" style=\"align:center;\"\u002F>\n\u003C\u002Fp>\n\n| 时间 | 方法 | 论文来源 | 任务 | 工具基模型 | 代理基模型 | 微调方法 |\n|------|--------|-------|---------|---------------|----------------|--------|\n| 2025.10 | QAgent | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2510.08383)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002FLivingFutureLab\u002FQAgent) | QA, RAG | Qwen2.5-3B | Qwen-7B | GRPO |\n| 2025.10 | AgentFlow | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2510.05592)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Flupantech\u002FAgentFlow) | 网络搜索、规划、推理、数学 | Qwen2.5-7B | Qwen2.5-7B | Flow-GRPO |\n| 2025.10 | Advisor Models | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2510.02453)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Faz1326\u002Fadvisor-models) | 数学、推理 | Qwen2.5-7B, Qwen3-8B | GPT-4o-Mini, GPT-5, Claude4-Sonnet, GPT-4.1-Mini | GRPO |\n| 2025.10 | AutoGraph-R1 | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2510.15339)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002FHKUST-KnowComp\u002FAutoGraph-R1) | 知识图谱构建、RAG | 知识图谱构建器（Qwen2.5-3B\u002F7B） | 冻结式RAG生成器（Qwen2.5-7B） | GRPO |\n| 2025.10 | MAE | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2510.23595)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fulab-uiuc\u002FMulti-agent-Evolve) | 数学、编程、常识推理 | Qwen2.5-3B | Qwen2.5-3B | REINFORCE++ |\n| 2025.09 | Mem-α | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2509.25911)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fwangyu-ustc\u002FMem-alpha) | 检索、测试时学习、长距离理解 | Qwen3-4B | Qwen3-4B, Qwen3-32B, GPT-4.1-Mini | GRPO |\n| 2025.08 | AI-SearchPlanner | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2508.20368) | 网络QA | Qwen3-32b | Qwen2.5-7B | PPO |\n| 2025.08 | Memento | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2508.16153)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002FAgent-on-the-Fly\u002FMemento) | 长期推理、网络研究、QA、学术推理 | Q函数（两层MLP） | GPT-4.1 | 软Q学习 |\n| 2025.08 | R-Zero | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2508.05004)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002FChengsong-Huang\u002FR-Zero) | 数学、推理 | Qwen3-4B, Qwen3-8B, OctoThinker-3B, OctoThinker-8B | Qwen3-4B, Qwen3-8B, OctoThinker-3B, OctoThinker-8B | GRPO |\n| 2025.06 | Sysformer | arXiv\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2506.15751) | QA、RAG | 小型Transformer | LLaMA-2-7B, LLaMA-3.1-8B, Mistral-7B, Phi-3.5-mini, Zephyr-7B-beta | 监督学习 |\n| 2025.05 | s3 | EMNLP'25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2505.14146)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fpat-jj\u002Fs3) | QA、RAG | Qwen2.5-7B | Qwen2.5-7B, Qwen2.5-14B, Claude-3-Haiku | PPO |\n| 2024.10 | Matryoshka Pilot | NeurIPS'25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2410.20749)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Flichangh20\u002FMatryoshka) | 数学、规划、推理 | LLaMA3-8B, Qwen2.5-7B | GPT-4o-Mini, GPT-3.5-Turbo | DPO, IDPO |\n| 2024.06 | CoBB | EMNLP'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2406.18695)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fbbuing9\u002FCoBB) | QA、数学 | Mistral-7b-inst-v2 | GPT-3.5-Turbo, Claude-3-Haiku, Phi-3-mini-4k-inst, Gemma-1.1-7B-it, Mistral-7B-inst-v2 | SFT, ORPO |\n| 2024.05 | Medadapter | EMNLP'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2405.03000)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fwshi83\u002FMedAdapter) | 医疗QA、自然语言推理、关系问答 | BERT-Base-Uncased | GPT-3.5-Turbo | SFT, BPO |\n| 2024.03 | BLADE | AAAI'25\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2403.18365)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002FCSHaitao\u002FBLADE) | 领域特定QA | BLOOMZ-1b7 | ChatGPT, ChatGLM, Baichuan, Qwen | SFT, BPO |\n| 2024.02 | ARL2 | ACL'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2402.13542)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fzhanglingxi-cs\u002FARL2) | QA | LLaMA2-7B | GPT-3.5-Turbo | 对比学习 |\n| 2024.02 | EVOR | EMNLP'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2402.12317)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fxlang-ai\u002FEVOR) | 基于RAG的编码 | GPT-3.5-Turbo | GPT-3.5-Turbo, CodeLLaMA | 提示工程 |\n| 2024.02 | Bbox-Adapter | ICML'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2402.08219)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fhaotiansun14\u002FBBox-Adapter) | QA | DeBERTa-v3-base (0.1B), DeBERTa-v3-large (0.3B) | GPT-3.5-Turbo, Mixtral-8x7B | 对比学习 |\n| 2024.01 | Proxy-Tuning | COLM'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2401.08565)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Falisawuffles\u002Fproxy-tuning) | QA、数学、代码 | LLaMA2-7B | LLaMA2-70B | 代理微调 |\n| 2024.01 | BGM | ACL'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2401.06954) | QA、个性化生成（NQ、HotpotQA、邮件、书籍） | T5-XXL-11B | PaLM2-S | SFT, PPO |\n| 2023.10 | RA-DIT | ICLR'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2310.01352) | 知识密集型任务（MMLU、NQ、TQA、ELI5、HotpotQA等） | DRAGON+ | LLaMA-65B | SFT, LSR |\n| 2023.06 | LLM-R | EACL'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2307.07164)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002FLMOps\u002Ftree\u002Fmain\u002Fllm_retriever) | 零样本NLU（阅读理解、QA、NLI、释义、情感分析、摘要生成） | E5-base | GPT-Neo-2.7B, LLaMA-13B, GPT-3.5-Turbo | 对比学习 |\n| 2023.05 | AAR | ACL'23\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2305.17331)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002FOpenMatch\u002FAugmentation-Adapted-Retriever) | 零样本泛化（MMLU、PopQA） | ANCE、Contriever | Flan-T5-Small、InstructGPT | 对比学习 |\n| 2023.05 | ToolkenGPT | NeurIPS'23\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2305.11554)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002FBer666\u002FToolkenGPT) | 数值推理、QA、计划生成 | Token Embedding | GPT-J 6B, OPT-6.7B, OPT-13B | 代理微调 |\n| 2023.03 | UPRISE | EMNLP'23\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Faclanthology.org\u002F2023.emnlp-main.758\u002F)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002FLMOps) | 零样本NLU（阅读理解、QA、NLI、释义、情感分析、摘要生成） | GPT-Neo-2.7B | BLOOM-7.1B, OPT-66B, GPT-3-175B | 对比学习 |\n| 2023.01 | REPLUG | NAACL'24\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_3dd1c217df3f.png\" width=\"16\" height=\"16\"> [论文](https:\u002F\u002Faclanthology.org\u002F2024.naacl-long.463.pdf)\u003Cbr>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_readme_1d4a1e4fac9a.png\" width=\"16\" height=\"16\"> [代码](https:\u002F\u002Fgithub.com\u002Fswj0419\u002FREPLUG) | QA | Contriever | GPT3-175B, PaLM, Codex, LLaMA-13B | 代理微调、LSR |\n\n---\n\n\n\n## 引用\n\n如果您觉得本仓库有用，请考虑引用我们的综述：\n\n```\n@article{jiang2025adaptation,\n  title={智能体AI的适应性},\n  author={蒋鹏程、林嘉诚、史志毅、王子峰、何陆曦、吴一辰、钟明、宋沛洋、张启正、王恒等},\n  journal={arXiv预印本 arXiv:2512.16301},\n  year={2025}\n}\n```\n\n## 贡献\n\n我们欢迎各位贡献！请随时提交Pull Request，以添加新论文或更新现有条目。\n\n---\n\n\u003Cp align=\"center\">\n  \u003Cbr>\n  \u003Csub>(ﾉ◕ヮ◕)ﾉ*:･ﾟ✧ 继续探索智能体AI的精彩世界吧！✧ﾟ･: *ヽ(◕ヮ◕ヽ)\u003C\u002Fsub>\n\u003C\u002Fp>","# Awesome-Adaptation-of-Agentic-AI 快速上手指南\n\n**项目简介**：\n`Awesome-Adaptation-of-Agentic-AI` 并非一个单一的代码库或工具框架，而是一个**精选论文与资源列表**。它系统性地整理了关于“智能体（Agentic AI）自适应策略”的最新研究成果，涵盖了基于强化学习（RL）、监督微调（SFT）和直接偏好优化（DPO）等多种方法。\n\n本指南旨在帮助开发者快速理解该列表的结构，并指导如何根据列表中的具体论文（如 Orion, Tool-R1, DeepRetrieval 等）找到对应的代码仓库进行复现和研究。\n\n## 环境准备\n\n由于本项目是资源索引，实际运行需针对列表中具体的论文项目进行环境配置。大多数列出的项目基于主流大模型框架，通用前置要求如下：\n\n*   **操作系统**: Linux (推荐 Ubuntu 20.04\u002F22.04) 或 macOS\n*   **Python 版本**: >= 3.9 (部分新项目可能需要 3.10+)\n*   **GPU**: 支持 CUDA 的 NVIDIA 显卡 (显存需求视具体模型而定，通常建议 24GB+ 以运行 7B-14B 参数模型)\n*   **核心依赖**:\n    *   PyTorch >= 2.0\n    *   Transformers >= 4.35\n    *   vLLM 或 TGI (用于高效推理)\n    *   Git\n\n> **国内加速建议**：\n> 在安装 Python 依赖时，推荐使用清华或阿里镜像源以提升下载速度：\n> ```bash\n> pip install -r requirements.txt -i https:\u002F\u002Fpypi.tuna.tsinghua.edu.cn\u002Fsimple\n> ```\n\n## 安装步骤\n\n由于本仓库本身不包含可执行的核心训练代码，\"安装\"过程实质上是**克隆目标论文的代码仓库**。请按照以下步骤操作：\n\n### 1. 克隆本资源列表仓库\n首先获取最新的论文列表和分类索引：\n\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002Fpat-jj\u002FAwesome-Adaptation-of-Agentic-AI.git\ncd Awesome-Adaptation-of-Agentic-AI\n```\n\n### 2. 选择并克隆具体项目\n浏览 `README.md` 中的表格，找到你感兴趣的方法（例如 `Tool-R1` 或 `DeepRetrieval`），点击表格中的 **[Code]** 链接跳转至对应 GitHub 仓库。\n\n以 **Tool-R1** 为例，安装其具体环境的命令通常如下（请在目标项目的根目录下执行）：\n\n```bash\n# 示例：克隆 Tool-R1 项目\ngit clone https:\u002F\u002Fgithub.com\u002FYBYBZhang\u002FTool-R1.git\ncd Tool-R1\n\n# 创建虚拟环境\npython -m venv venv\nsource venv\u002Fbin\u002Factivate  # Windows 用户使用: venv\\Scripts\\activate\n\n# 安装依赖 (优先使用国内镜像)\npip install -r requirements.txt -i https:\u002F\u002Fpypi.tuna.tsinghua.edu.cn\u002Fsimple\n```\n\n*注意：不同论文项目的依赖名称可能不同（如 `requirements.txt`, `setup.py` 或 `pyproject.toml`），请以具体项目说明为准。*\n\n## 基本使用\n\n本资源库的核心价值在于**导航**。使用流程分为“检索策略”和“复现实验”两步。\n\n### 1. 检索自适应策略\n打开本地或在线的 `README.md`，根据任务需求查找对应类别：\n\n*   **需要提升工具调用能力？**\n    *   查看 `Agent Adaptation` -> `A1: Tool Execution Signaled` 章节。\n    *   关注基于 RL 的方法（如 `Tool-R1`, `Orion`）或 SFT 方法（如 `ToolLLM`）。\n*   **需要优化特定领域推理（如数学证明、代码生成）？**\n    *   查看 `Task(s)` 列，筛选 \"Formal Theorem Proving\" 或 \"Coding\"。\n    *   推荐尝试 `AlphaProof`, `DeepSeek-Prover-V2`, 或 `Code-R1`。\n*   **需要多轮路由或检索增强？**\n    *   参考 `Router-R1` 或 `DeepRetrieval`。\n\n### 2. 运行示例（以通用结构为例）\n一旦进入具体项目仓库（例如 `Tool-R1`），典型的使用流程如下：\n\n**步骤 A: 数据准备**\n大多数项目需要提供训练数据或测试集。\n```bash\n# 示例：下载数据集\npython scripts\u002Fdownload_data.py --dataset tool_bench\n```\n\n**步骤 B: 模型微调\u002F推理**\n根据项目提供的脚本启动训练或推理。以下是基于强化学习（GRPO\u002FPPO）的典型启动命令：\n\n```bash\n# 示例：启动 GRPO 训练 (具体参数需参考该项目文档)\naccelerate launch --num_processes 8 train_grpo.py \\\n    --model_name_or_path Qwen\u002FQwen2.5-7B-Instruct \\\n    --dataset_path data\u002Ftool_usage.jsonl \\\n    --output_dir outputs\u002Ftool-r1-qwen2.5\n```\n\n**步骤 C: 评估结果**\n```bash\n# 示例：运行评估脚本\npython eval.py --checkpoint outputs\u002Ftool-r1-qwen2.5\u002Fcheckpoint-1000\n```\n\n### 3. 引用研究\n如果您在研究中使用了该列表整理的内容，请在论文中引用其配套文章：\n\n```bibtex\n@article{jiang2025adaptation,\n  title={Adaptation of Agentic AI},\n  author={Jiang, Pengcheng and Lin, Jiacheng and Shi, Zhiyi and Wang, Zifeng and He, Luxi and Wu, Yichen and Zhong, Ming and Song, Peiyang and Zhang, Qizheng and Wang, Heng and others},\n  journal={arXiv preprint arXiv:2512.16301},\n  year={2025}\n}\n```","某金融科技团队正致力于构建一个能自动执行复杂量化策略并调用多种外部数据 API 的智能代理系统，以应对瞬息万变的股市行情。\n\n### 没有 Awesome-Adaptation-of-Agentic-AI 时\n- **工具调用僵化**：当新增或更新金融数据接口时，智能代理无法自适应调整调用逻辑，必须人工重写大量代码才能兼容。\n- **错误恢复能力弱**：一旦某个 API 返回异常格式或非预期数据，代理往往直接崩溃或陷入死循环，缺乏基于执行信号的自我修正机制。\n- **研发效率低下**：团队需从零摸索如何让大模型学会使用新工具，缺乏如 Tool-R1 或 Orion 等经过验证的强化学习微调策略参考，试错成本极高。\n- **场景适配困难**：面对从“文档 OCR 识别”到“形式化定理证明”等不同任务，难以快速找到对应的代理适应方案，导致项目延期。\n\n### 使用 Awesome-Adaptation-of-Agentic-AI 后\n- **动态自适应执行**：借鉴列表中“工具执行信号驱动”的策略，代理能根据 API 返回的实时状态自动调整后续操作，无需人工干预即可处理接口变更。\n- **鲁棒性显著提升**：利用成熟的 RL-based 方法（如 GRPO 微调），代理在遇到数据异常时能自主尝试备选路径，大幅降低系统崩溃率。\n- **加速落地进程**：团队直接复用仓库中针对特定任务（如 WebGen-Agent 或 AlphaProof）的最佳实践论文与代码，将新工具集成周期从数周缩短至几天。\n- **精准策略匹配**：通过清晰的分类索引，开发者能迅速定位适合当前金融场景的“代理监督式工具适应”方案，实现技术选型的最优化。\n\nAwesome-Adaptation-of-Agentic-AI 通过汇聚前沿的代理适应策略，将智能系统从“静态脚本执行者”进化为具备自我演进能力的“动态决策专家”。","https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fpat-jj_Awesome-Adaptation-of-Agentic-AI_3dd1c217.png","pat-jj","Patrick Jiang","https:\u002F\u002Foss.gittoolsai.com\u002Favatars\u002Fpat-jj_46fc13af.png","cs phd at uiuc","UIUC","Champaign, IL",null,"patpcj","https:\u002F\u002Fpat-jj.github.io\u002F","https:\u002F\u002Fgithub.com\u002Fpat-jj",633,55,"2026-04-16T04:12:03","NOASSERTION","","未说明",{"notes":91,"python":89,"dependencies":92},"该仓库（Awesome-Adaptation-of-Agentic-AI）是一个 curated list（精选列表），主要用于整理和展示关于代理式 AI（Agentic AI）自适应策略的学术论文及相关代码链接，其本身不是一个可直接运行的软件工具或框架。因此，README 中未包含具体的操作系统、GPU、内存、Python 版本或依赖库等运行环境需求。如需运行列表中提及的具体项目（如 Orion, Tool-R1, DeepSeek-Prover 等），请访问各项目的独立代码仓库链接以获取相应的环境配置说明。",[],[13,35],[95,96,97],"adaptation","agentic-ai","large-language-models","2026-03-27T02:49:30.150509","2026-04-17T09:54:12.835312",[],[102],{"id":103,"version":104,"summary_zh":105,"released_at":106},297337,"v0.5","预发布版本。","2025-12-09T14:40:12"]