📈 Benchmark Statistics

31

Total Memes

151

Total Evaluations


51

Unique Contributors

20

Registered Users


About This Benchmark

This crowd-sourced benchmark gathers human evaluations of memes to advance AI systems' comprehension of humor, emotional nuance, cultural references, and visual-semantic alignment. For each meme, participants can:

  • Submit their own textual descriptions
  • Edit or refine existing descriptions
  • Upvote or downvote descriptions based on accuracy and insight
  • Classify the meme's humor type (e.g., absurdity, irony, wordplay, observational, etc.)
  • Identify primary emotions conveyed
  • Provide detailed annotations of visual elements and contextual references

Contribute Now