Covert Bias: The Severity of Social Views' Unalignment in Language Models Towards Implicit and Explicit Opinion

Documenti analoghi

• EMBRACE: Shaping Inclusive Opinion Representation by Aligning Implicit Conversations with Social Norms
  di: Aldayel, Abeer, et al.
  Pubblicazione: (2025)
• Measuring Implicit Bias in Explicitly Unbiased Large Language Models
  di: Bai, Xuechunzi, et al.
  Pubblicazione: (2024)
• Hatred Stems from Ignorance! Distillation of the Persuasion Modes in Countering Conversational Hate Speech
  di: Alyahya, Ghadi, et al.
  Pubblicazione: (2024)
• Inference-Time Reasoning Selectively Reduces Implicit Social Bias in Large Language Models
  di: Apsel, Molly, et al.
  Pubblicazione: (2026)
• Do Political Opinions Transfer Between Western Languages? An Analysis of Unaligned and Aligned Multilingual LLMs
  di: Weeber, Franziska, et al.
  Pubblicazione: (2025)