HumorRank ranks nine LLMs by humor generation using tournament-style pairwise LLM judgments, producing cross-judge-stable leaderboards where a 7B specialized model beats much larger models.
Evaluating humor generation in an improvisational comedy setting
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
HumorRank: A Tournament-Based Leaderboard for Evaluating Humor Generation in Large Language Models
HumorRank ranks nine LLMs by humor generation using tournament-style pairwise LLM judgments, producing cross-judge-stable leaderboards where a 7B specialized model beats much larger models.