← All works
Research paper

Chatbot Arena: An Open Platform for Evaluating LLMs by Human Preference

At a glance

Presented Chatbot Arena, comparing language models through people's votes on anonymous answers.

In one paragraph

Presented Chatbot Arena, an open platform where people compare anonymous chatbot answers and vote for their preferred response. Analyzed more than 240,000 votes to test how reliably these comparisons rank language models.

Authors 11

  1. Wei-Lin Chiang
  2. Lianmin Zheng
  3. Ying Sheng
  4. Anastasios Angelopoulos
  5. Tianle (Tim) Li
  6. Dacheng Li
  7. Hao Zhang
  8. Banghua Zhu
  9. Michael I. Jordan
  10. Joseph Gonzalez
  11. Ion Stoica

Organizations

Credited for this work at publication.