Jinhao Jiang

Research Interest

研究兴趣

My research interest focuses on General Autonomous Agent with an emphasis on Data Scaling and RL Training of Coding Agent and General Agent. Currently, I am a core contributor to the release of the Seed language models (history includes Seed1.8 and Seed2.0) in ByteDance, with responsibility for instruction following and generalization capabilities.

我的研究兴趣集中在通用自主智能体，特别关注 代码智能体和通用智能体的数据扩展和强化学习训练。目前，我在字节跳动核心参与Seed语言模型（历史版本包括Seed1.8和Seed2.0）的合版发布，负责指令遵循和通用泛化能力。

If there are suitable post-training positions for LLMs in the industry, especially job openings related to code agents and general-purpose agents, feel free to reach out to me via email. 如果在工业界有合适的大语言模型后训练，尤其是代码智能体和通用智能体方向的工作机会，如果对我有兴趣，欢迎通过邮件联系我。

News

新闻

[2026-06-12] Seed 2.0 Lite (0428) is released, delivering substantial improvements in coding capabilities and full-modal performance compared to version 0215.
[2026-02-14] Seed 2.0 is released. Welcome to check the model card and blog for more use cases.
[2026-02-04] We release SWE-Master and SWE-World, which aim to democratize the training of code agents.
[2025-12-18] Seed 1.8 is released. Welcome to check the model card and blog for more use cases. I am responsible for the general search agent ability of Seed 1.8 (BrowseComp 67.6, BrowseComp-ZH 81.3, GAIA 87.4, HLE 40.9), and have achieved more efficient search efficiency than DeepSeek.
[2025-08-22] We present S1-Search, which achieves SOTA performance across a range of deep search benchmarks using open-source LLMs.
[2025-08-21] I have five papers accepted by EMNLP 2025, including CAFE (Main), StickerTTS (Main), ManuSearch (Findings), R1-Searcher++ (Findings), and SimpleDeepSearcher (Findings), congratulations to my co-authors!
[2025-05-23] We release ManuSearch, which aims to push the LLM-based AI search with a curated complex search benchmark and a strong training-free multi-agent framework.
[2025-05-22] We release SimpleDeepSearcher and R1-Searcher++, which are the data engineering method for collecting small but refined SFT dataa, and a novel RL framework for incentivizing the dynamic knowledge acquisition of LLMs from internal knowledge and external tools.
[2025-05-21] I have four papers accepted by ACL 2025 main conference, including YuLan-Mini, Llama-3-SynE, KG-Agent, and LongReD, congratulations to my co-authors!
[2025-04-29] We release RV-Syn, which is a rational and verifiable mathematical data synthesis method based on structured function library. It has achieved more efficent scaling curve compared to latest methodes, such as ScaleQuest, NuminaMath.
[2025-04-22] I'll present my work "Mix-CPT: A Domain Adaptation Framework via Decoupling Knowledge Learning and Format Alignment" poster at ICLR-2025 in Singapore. It's on April 26th, from 15:00 - 17:30 in Hall 3 + Hall 2B. Welcome to have a chat!
[2025-03-07] We release R1-Searcher, which is the first technical report to apply the RL of the R1 paradigm to the RAG scenario. It has achieved significant performance improvements across multiple evaluation datasets, marking an important step towards Deep Research!

[2026-06-12] Seed 2.0 Lite (0428) 发布，相比版本 0215，在代码能力和全模态性能上取得了显著提升。
[2026-02-14] Seed 2.0 发布。欢迎查看 model card 和博客获取更多信息。
[2026-02-04] 我们发布了 SWE-Master 和 SWE-World, 其目的是使代码智能体的训练平民化。
[2025-12-18] Seed 1.8 发布。欢迎查看 model card 和 blog 获取更多信息. 我负责 Seed 1.8 的通用搜索功能 (BrowseComp 67.6, BrowseComp-ZH 81.3, GAIA 87.4, HLE 40.9)，同时实现了比 DeepSeek 更高效的搜索效率。
[2025-08-22] 我们提出了 S1-Search, 它使用开源的大型语言模型（LLMs），在一系列深度搜索基准测试中取得了最先进（SOTA）的性能。
[2025-08-21] 我有五篇论文被EMNLP 2025主会议接收，包括 CAFE (Main), StickerTTS (Main), ManuSearch (Findings), R1-Searcher++ (Findings), and SimpleDeepSearcher (Findings), 祝贺我的合作者！
[2025-05-23] 我们发布了 ManuSearch，旨在推动基于大语言模型的AI搜索，并提供了一个精心策划的复杂搜索基准和一个强大的通用免训的多智能体框架。
[2025-05-22] 我们发布了 SimpleDeepSearcher 和 R1-Searcher++，它们是收集少而精炼的SFT数据和激励LLM从内部知识和外部工具中动态获取知识的新方法。
[2025-05-21] 我有四篇论文被ACL 2025主会议接收，包括 YuLan-Mini、Llama-3-SynE、KG-Agent和LongReD，祝贺我的合作者！
[2025-04-29] 我们发布了 RV-Syn，它是一种基于结构化函数库的理性且可验证的数学数据合成方法。与最新的方法相比，它具有更高效的缩放曲线，如ScaleQuest和NuminaMath。
[2025-04-22] 我将在新加坡的ICLR-2025会议上展示我的工作 "Mix-CPT: 通过解耦知识学习和格式对齐的领域适应框架"。展示时间是4月26日下午15:00-17:30，地点在3号厅+2B厅。欢迎来交流！
[2025-03-07] 我们发布了 R1-Searcher，这是首个将R1范式的强化学习应用于RAG场景的技术报告。它在多个评估数据集上取得了显著的性能提升，标志着向深度研究迈出了重要一步！

Selected Publications

精选论文

(* indicates equal contribution, † indicates corresponding author)

(* 表示共同一作, † 表示通讯作者)

Here, I've listed my work as a (co-) first author. For the complete list of my publications, please visit my Google Scholar profile.

这里列出了我作为（共同）第一作者的工作。有关我发表的完整论文列表，请访问我的谷歌学术主页。

ManuSearch: Democratizing Deep Search in Large Language Models with a Transparent and Open Multi-Agent Framework

Lisheng Huang*, Yichen Liu*, Jinhao Jiang*, Rongxiang Zhang, Jiahao Yan, Junyi Li, Wayne Xin Zhao

EMNLP-Findings, 2025
R1-Searcher++: Incentivizing the Dynamic Knowledge Acquisition of LLMs via Reinforcement Learning

Huatong Song*, Jinhao Jiang*, Wenqing Tian, Zhipeng Chen, Yuhuan Wu, Jiahao Zhao, Yingqian Min, Wayne Xin Zhao, Lei Fang, Ji-Rong Wen

EMNLP-Findings, 2025
R1-Searcher: Incentivizing the Search Capability in LLMs via Reinforcement Learning

Huatong Song*, Jinhao Jiang*, Yingqian Min, Jie Chen, Zhipeng Chen, Wayne Xin Zhao†, Lei Fang, Ji-Rong Wen

Technical Report, 2025

RV-Syn: Rational and Verifiable Mathematical Reasoning Data Synthesis based on Structured Function library

Jiangpeng Wang*, Jinhao Jiang*, Zhiqiang Zhang, Jun Zhou, Wayne Xin Zhao†

arXiv, 2025

CAFE: Retrieval Head-based Coarse-to-Fine Information Seeking to Enhance Multi-document Question Answering Capability

Han Peng*, Jinhao Jiang*, Zican Dong*, Wayne Xin Zhao†, Lei Fang

EMNLP, 2025

Imitate, explore, and self-improve: A reproduction report on slow-thinking reasoning systems

Yingqian Min*, Zhipeng Chen*, Jinhao Jiang*, Jie Chen, Jia Deng, Yiwen Hu, Yiru Tang, Jiapeng Wang, Xiaoxue Cheng, Huatong Song, Wayne Xin Zhao†, Zheng Liu, Zhongyuan Wang, Ji-Rong Wen

Technical Report, 2025

Enhancing LLM Reasoning with Reward-guided Tree Search

Jinhao Jiang*, Zhipeng Chen*, Yingqian Min*, Jie Chen, Xiaoxue Cheng, Jiapeng Wang, Yiru Tang, Haoxiang Sun, Jia Deng, Wayne Xin Zhao†, Zheng Liu, Dong Yan, Jian Xie, Zhongyuan Wang, Ji-Rong Wen

Technical Report, 2024

Mix-CPT: A Domain Adaptation Framework via Decoupling Knowledge Learning and Format Alignment

Jinhao Jiang*, Junyi Li*, Wayne Xin Zhao†, Yang Song, Tao Zhang, Ji-Rong Wen

International Conference on Learning Representations (ICLR), 2025

RAG-Star: Enhancing Deliberative Reasoning with Retrieval Augmented Verification and Refinement

Jinhao Jiang*, Jiayi Chen*, Junyi Li*, Ruiyang Ren, Shijie Wang, Wayne Xin Zhao†, Yang Song, Tao Zhang

The North American Chapter of the Association for Computational Linguistics (NAACL), 2025

KG-Agent: An Efficient Autonomous Agent Framework for Complex Reasoning over Knowledge Graph

Jinhao Jiang*, Kun Zhou*, Wayne Xin Zhao†, Yang Song, Chen Zhu, Hengshu Zhu, Ji-Rong Wen

The 63rd Annual Meeting of the Association for Computational Linguistics (ACL), 2025

StructGPT: A General Framework for Large Language Model to Reason over Structured Data

Jinhao Jiang*, Kun Zhou*, Zican Dong, Keming Ye, Wayne Xin Zhao†, Ji-Rong Wen

The 2023 Conference on Empirical Methods in Natural Language Processing (EMNLP), 2023

ReasoningLM: Enabling Structural Subgraph Reasoning in Pre-trained Language Models for Question Answering over Knowledge Graph

Jinhao Jiang, Kun Zhou, Wayne Xin Zhao†, Yaliang Li, Ji-Rong Wen

The 2023 Conference on Empirical Methods in Natural Language Processing (EMNLP), 2023

UniKGQA: Unified Retrieval and Reasoning for Solving Multi-hop Question Answering Over Knowledge Graph

Jinhao Jiang*, Kun Zhou*, Wayne Xin Zhao†, Ji-Rong Wen

International Conference on Learning Representations (ICLR), 2023

Great Truths are Always Simple: A Rather Simple Knowledge Encoder for Enhancing the Commonsense Reasoning Capacity of Pre-Trained Models

Jinhao Jiang*, Kun Zhou*, Wayne Xin Zhao†, Ji-Rong Wen

The North American Chapter of the Association for Computational Linguistics (NAACL-Findings), 2022

Complex Knowledge Base Question Answering: A Survey

Yunshi Lan*, Gaole He*, Jinhao Jiang, Jing Jiang, Wayne Xin Zhao†, Ji-Rong Wen

IEEE Transactions on Knowledge and Data Engineering (TKDE), 2022

Grants

奖项

2021 Outstanding Graduates of Sichuan Province (winning ratio 3.7%), Education Department of Sichuan.
2020 China National Scholarship (top 1.5%), Ministry of Education of the People's Republic of China.
2019 China National Scholarship (top 1.5%), Ministry of Education of the People's Republic of China.

2021 四川省优秀毕业生（获奖比例3.7%），四川省教育厅。
2020 国家奖学金（前1.5%），中华人民共和国教育部。
2019 国家奖学金（前1.5%），中华人民共和国教育部。

Professional Service

学术服务

Journal: TALLIP, Computational Intelligence, Information Retrieval Journa
Conference: ICLR, NIPS, ACL, EMNLP

期刊: TALLIP, Computational Intelligence, Information Retrieval Journal
会议: ICLR, NIPS, ACL, EMNLP

About Me [GitHub] [Google Scholar]

关于我 [GitHub] [Google Scholar]