LLM Validation System

Help validate AI-generated annotations of social media content for harmful language patterns.

Please sign in or register to begin annotating tweets.

Sign In Register

About This Project

This system collects human validation of AI-generated annotations for Greek social media content. We're analyzing posts for:

  • Dehumanizing Language - Animal metaphors and subhuman characterizations
  • Victimhood Narratives - Claims of Greeks being victims of outgroups
  • Hate Speech - Direct expressions of hatred or hostility
  • Violent Speech - Calls for violence or violent action
  • Sympathy/Support - Expressions of support for outgroups

Sample: 300 tweets, stratified by targeted group and LLM score patterns