LLM Agent Training

LLM: Large Language Model

LLM agent training studies how to make language models act reliably over multiple steps, using tools, software, websites, or simulated worlds to complete tasks. Common approaches combine supervised fine-tuning on interaction traces with reinforcement learning from execution-verified outcomes, alongside distillation, preference learning, and feedback on individual steps. Researchers build executable environments and benchmarks such as SWE-bench, WebArena, Terminal-Bench, and τ²-Bench to provide repeatable practice and test tool use, coding, search, and long-horizon workflows. A central challenge is transferring gains beyond training environments while balancing task success with efficiency, safety, and robustness to changing tools and users.

Latest papers

590 papers130 papers in the last four weeks, up 132% on the four weeks before. 0.8% of all new papers.

Subliminal Transfer of Unsafe Behaviors in AI Agent Distillation
Apr 16, 2026Jacob Dang, Brian Y. Xie, Omar G. YounisSubliminal Learning Language Model Distillation
UCLA · Santa Monica College · Mila, Silverstream AI
Does RL Expand the Capability Boundary of LLM Agents? A PASS@(k,T) Analysis
Apr 16, 2026Zhiyuan Zhai, Wenjing Yan, Xiaodan Shao +1LLM Agent Training RL for LLM Agents
Fudan University · Chinese University of Hong Kong · University of Waterloo
Mind DeepResearch Technical Report
Apr 16, 2026MindDR Team, Li Auto IncMulti-Agent LLM Systems Deep Research Agents
MindDR Team, Li Auto Inc
TREX: Automating LLM Fine-tuning via Agent-Driven Tree-based Exploration
Apr 15, 2026Zerun Ma, Guoqiang Wang, Xinchen Xie +7LLM Fine-Tuning LLM Agent Training
Shanghai AI Laboratory · Fudan University
ToolOmni: Enabling Open-World Tool Use via Agentic learning with Proactive Retrieval and Grounded Execution
Apr 15, 2026Shouzheng Huang, Meishan Zhang, Baotian Hu +1LLM Tool Use LLM Agent Training
Harbin Institute of Technology (Shenzhen)
Towards Scalable Lightweight GUI Agents via Multi-role Orchestration
Apr 15, 2026Ziwei Wang, Junjie Zheng, Leyang Yang +7GUI Agents LLM Agent Training
Zhejiang Key Laboratory of Accessible Perception and Intelligent Systems, Zhejiang University · College of Computer Science and Technology, Zhejiang University · AntGroup
QuarkMedSearch: A Long-Horizon Deep Search Agent for Exploring Medical Intelligence
Apr 14, 2026Zhichao Lin, Zhichao Liang, Gaoqiang Liu +6Agentic Search LLM Agent Training
Qwen Applications Business Group, Alibaba · Shanghai Jiao Tong University
From Imitation to Discrimination: Progressive Curriculum Learning for Robust Web Navigation
Apr 14, 2026Chuang Peng, Wei Zhang, Renshuai Tao +2Web Agents LLM Agent Training
Beijing Jiaotong University · Beihang University
UniToolCall: Unifying Tool-Use Representation, Data, and Evaluation for LLM Agents
Apr 13, 2026Yijuan Liang, Xinghao Chen, Yifan Ge +5LLM Agent Evaluation AI Agent Benchmarks
University of Science and Technology of China · Zhejiang Key Laboratory of Industrial Intelligence and Digital Twin, Eastern Institute of Technology, Ningbo · Department of Computing, The Hong Kong Polytechnic University
YIELD: A Large-Scale Dataset and Evaluation Framework for Information Elicitation Agents
Apr 13, 2026Victor De Lima, Grace Hui YangOutput Elicitation LLM Agent Evaluation
InfoSense Lab, Georgetown University
HTAA: Enhancing LLM Planning via Hybrid Toolset Agentization & Adaptation
Apr 13, 2026Chengrui Huang, Junshuo Zhang, Zhiyuan Ma +7Tool-Use Planning LLM Tool Use
University of Electronic Science and Technology of China · 2DiDi Global Inc.
Pioneer Agent: Continual Improvement of Small Language Models in Production
Apr 10, 2026Dhruv Atreja, Julia White, Nikhil Nayak +5LLM Agent Training LLM Fine-Tuning
Fastino Labs
E3-TIR: Enhanced Experience Exploitation for Tool-Integrated Reasoning
Apr 10, 2026Weiyang Guo, Zesheng Shi, Liye Zhao +5Tool-Integrated Reasoning LLM Agent Training
Harbin Institute of Technology, Shenzhen, China · Huawei Technologies Co., Ltd. · Hong Kong University of Science and Technology
CONSCIENTIA: Can LLM Agents Learn to Strategize? Emergent Deception and Trust in a Multi-Agent NYC Simulation
Apr 10, 2026Aarush Sinha, Arion Das, Soumyadeep Nag +7Strategic Reasoning Risks in Language Models Emergent Behavior in Multi-Agent Systems
University of Copenhagen · IIIT Ranchi · ISI Kolkata +6
Training Language Models for Bilateral Trade with Private Information
Apr 10, 2026Dirk Bergemann, Soheil Ghili, Xinyang Hu +2Multi-Agent Negotiation LLM Agent Training
Department of Economics, Yale University · Yale School of Management, Yale University · Department of Statistics and Data Science, Yale University +1
MolmoWeb: Open Visual Web Agent and Open Data for the Open Web
Apr 9, 2026Tanmay Gupta, Piper Wolters, Zixian Ma +13Web Agents LLM Agent Training
Allen Institute for AI · University of Washington · UNC-Chapel Hill
Awakening the Sleeping Agent: Lean-Specific Agentic Data Reactivates General Tool Use in Goedel Prover
Apr 9, 2026Jui-Hui Chung, Hongzhou Lin, Lai Jiang +2LLM Tool Use LLM Post-Training
Princeton Language and Intelligence · Amazon · Department of Computer Science, University of Illinois at Urbana-Champaign
TOOLCAD: Exploring Tool-Using Large Language Models in Text-to-CAD Generation with Reinforcement Learning
Apr 9, 2026Yifei Gong, Xing Wu, Wenda Liu +1Text-to-CAD Generation LLM Tool Use
School of Computer Engineering & Science, Shanghai University
Structured Distillation of Web Agent Capabilities Enables Generalization
Apr 9, 2026Xing Han Lù, Siva ReddyLLM Agent Training Web Agents
Mila – Quebec AI Institute McGill University
CLEAR: Context Augmentation from Contrastive Learning of Experience via Agentic Reflection
Apr 8, 2026Linbo Liu, Guande Wu, Han Ding +7LLM Context Management LLM Agents
AWS AI
AgentGate: A Lightweight Structured Routing Engine for the Internet of Agents
Apr 8, 2026Yujun Cheng, Enfang Cui, Hao Qin +2LLM Agent Routing Structured Prediction
School of Artificial Intelligence, University of Science and Technology Beijing, Beijing, China · China Telecom Research Institute, China · Hangzhou Institute for Advanced Study, University of Chinese Academy of Sciences, Hangzhou, China
Synthetic Sandbox for Training Machine Learning Engineering Agents
Apr 6, 2026Yuhang Zhou, Lizhu Zhang, Yifan Wu +4LLM Agent Training RL for LLM Agents
Meta AI
Search, Do not Guess: Teaching Small Language Models to Be Effective Search Agents
Apr 6, 2026Yizhou Liu, Qi Sun, Yulin Chen +2Agentic Search Small Language Models
University of Illinois Urbana-Champaign · New York University · Nanyang Technological University
Implementing surrogate goals for safer bargaining in LLM-based agents
Apr 6, 2026Caspar Oesterheld, Maxime Riché, Filip Sondej +2AI Agent Safety LLM Agent Safety
Squeez: Task-Conditioned Tool-Output Pruning for Coding Agents
Apr 4, 2026Ádám KovácsAI Coding Agents LLM Agent Training
KR Labs
Improving Role Consistency in Multi-Agent Collaboration via Quantitative Role Clarity
Apr 3, 2026Guoling Zhou, Wenpei Han, Fengqin Yang +3Behavioral Consistency of LLM Agents LLM Agents
School of Information Science and Technology, Northeast Normal University, Changchun, Jilin · School of Computer Science and Engineering, Guangxi Normal University, Guilin, Guangxi
From SWE-ZERO to SWE-HERO: Execution-free to Execution-based Fine-tuning for Software Engineering Agents
Apr 2, 2026Nikolai Ludwig, Wasi Uddin Ahmad, Somshubra Majumdar +1Software Engineering Agents LLM Agent Training
NVIDIA Santa Clara, CA 95051, USA
ORBIT: Scalable and Verifiable Data Generation for Search Agents on a Tight Budget
Apr 1, 2026Nandan Thakur, Zijian Chen, Xueguang Ma +1Synthetic Data Generation for Language Models LLM Agent Training
David R. Cheriton School of Computer Science University of Waterloo
Agentic Tool Use in Large Language Models
Apr 1, 2026Jinchao Hu, Meizhi Zhong, Kehai Chen +2LLM Agent Evaluation Tool-Using Agents
School of Computer Science and Technology, Harbin Institute of Technology, Shenzhen, China · Independent Researcher, China
Meta-TTL: Meta-Learning Self-Improvement Policies for Language Agents
Apr 1, 2026Zhanzhi Lou, Hui Chen, Yibo Li +2Meta-Learning Test-Time Training
National University of Singapore

Apr 16, 2026J/K move · Enter open · S save

Subliminal Transfer of Unsafe Behaviors in AI Agent Distillation

Jacob Dang, Brian Y. Xie, Omar G. Younis

UCLA · Santa Monica College · Mila, Silverstream AI

Subliminal Learning Language Model Distillation LLM Agent Training LLM Agent Safety Teacher-Student Learning

Open paper AskPDF arXiv

Recent work on subliminal learning demonstrates that language models can transmit semantic traits through data that is semantically unrelated to those traits. However, it remains unclear whether behavioral traits can transfer in agentic systems, where policies are learned from trajectories rather than static text. In this work, we provide the first empirical evidence that unsafe agent behaviors can transfer subliminally through model distillation across two complementary experimental settings. In our primary setting, we construct a teacher agent exhibiting a strong deletion bias, a tendency to perform destructive file-system actions via an API-style tool interface, and distill it into a student using only trajectories from ostensibly safe tasks, with all explicit deletion keywords rigorously filtered. In our secondary setting, we replicate the threat model in a native Bash environment, replacing API tool calls with shell commands and operationalizing the bias as a preference for issuing chmod as the first permission-related command over semantically equivalent alternatives such as chown or setfacl. Despite full keyword sanitation in both settings, students inherit measurable behavioral biases. In the API setting the student's deletion rate reaches 100% (versus a 5% baseline) under homogeneous distillation; in the Bash setting the student's chmod-first rate reaches 30%-55% (versus a 0%-10% baseline), with the strongest transfer observed in large-to-small distillation. Our results demonstrate that explicit data sanitation is an insufficient defense, and behavioral biases are encoded implicitly in trajectory dynamics regardless of the tool interface.