Tianle Li


I love ML and LLM

I love gym, philosophy, and math

EECS undergraduate student at UC Berkeley. Researcher at Berkeley Sky Computing Lab and a member of Large Model Systems Organization (LMSYS), advised by Professor Ion Stoica and Professor Joseph E. Gonzalez. I am part of the Chatbot Arena team, working on Large Language Model Evaluation.

Currently, I am also a Research Intern at Nexusflow.ai working on training Large Language Models (Starling and Raven), with Professor Jiantao Jiao. Previously, I was a software engineering intern at AMD, working on the Visti Model Composer library.

I am facsinated by Large Models and their capabilities in various domains, especially natural language. I am also very passionate about application of statistics and reinforcement learning theory in this domain.

selected publications

  1. arena-hard-img.png
    From Live Data to High-Quality Benchmarks: The Arena-Hard Pipeline
    Tianle Li*, Wei-Lin Chiang*, Evan Frick, Lisa Dunlap, Banghua Zhu, Joseph E. Gonzalez, Ion Stoica
    Apr 2024
  2. arena_log.png
    Chatbot Arena: An Open Platform for Evaluating LLMs by Human Preference
    Wei-Lin Chiang , Lianmin Zheng , Ying Sheng , and 8 more authors
    ICML 2024, Mar 2024
  3. lmsys-chat-1m-logo.png
    LMSYS-Chat-1M: A Large-Scale Real-World LLM Conversation Dataset
    Lianmin Zheng , Wei-Lin Chiang , Ying Sheng , and 10 more authors
    ICLR 2024 Spotlight, Mar 2024