benchmark · first seen 10 Nov 2026
human preference evaluation of chatbots
Also: Chatbot Arena, pairwise human voting, arena-style model evaluation
Ranking language model chatbots by crowdsourced blind pairwise human votes on their responses.
Mentions per week
PapersCodeDiscussion
Earliest sources
- AI Isn't Getting Dumber, We're Getting Pickier: Rising Standards in Human Evaluation of Large Language ModelsPapers · 10 Nov 2026 · Zihan Chen, Lei Nico Zheng, Di Zhu
- Leaderboard Hacking: Preference-Based Model Evaluations are Vulnerable to ManipulationPapers · 10 Nov 2026 · Eve Fleisig, Joachim Baumann, Dylan Hadfield-Menell, Dirk Hovy
Latest sources
- Leaderboard Hacking: Preference-Based Model Evaluations are Vulnerable to ManipulationPapers · 10 Nov 2026 · Eve Fleisig, Joachim Baumann, Dylan Hadfield-Menell, Dirk Hovy
- AI Isn't Getting Dumber, We're Getting Pickier: Rising Standards in Human Evaluation of Large Language ModelsPapers · 10 Nov 2026 · Zihan Chen, Lei Nico Zheng, Di Zhu