Kunlun Zhu
Scholar

Kunlun Zhu

Google Scholar ID: A5L7wZAAAAAJ
University of Illinois at Urbana-Champaign
Large Language ModelsFoundation AgentsAgents for ScienceAgents Safety
Citations & Impact
All-time
Citations
1,914
 
H-index
12
 
i10-index
13
 
Publications
17
 
Co-authors
9
list available
Resume (English only)
Academic Achievements
  • Paper "AI Scientist Safety" accepted by Nature Communications (Sep 2025)
  • Paper "SafeScientist" accepted at EMNLP 2025 Main Conference (Aug 2025)
  • Invited talk on "OpenManus Agent" at AMD Annual AI Conference (Jun 2025)
  • Published multiple top-tier conference papers including ACL 2025, ICML 2025, ICLR 2025, EMNLP Demo 2025, etc.
  • Co-developed TinyScientist: a lightweight framework for research agents
  • Built MultiAgentBench: a benchmark for evaluating collaboration and competition of LLM agents
  • Proposed RAGEval: a scenario-specific RAG evaluation dataset generation framework
Research Experience
  • Research Assistant at Tsinghua University's NLP group for two years, advised by Prof. Zhiyuan Liu
  • Algorithm Engineer at ModelBest Inc., advised by Prof. Zhiyuan Liu
  • Visiting student at Stanford University in summer 2025, collaborating with Prof. James Zou
  • 6-month Research Assistant at Mila Quebec AI Institute’s Graph Team, collaborating with Prof. Jian Tang
  • 1-year research intern at Carnegie Mellon University’s Robotics Institute, working with Prof. Katia Sycara
  • Member of OpenBMB