Confident, Calibrated, or Complicit: Safety Alignment and Ideological Bias in LLM Hate Speech Detection.

Sanjeevan Selvaganapathy, Mehwish Nasim

Browse the full ACL paper archive.