Paul Röttger

Latest

• The PRISM Alignment Dataset: What Participatory, Representative and Individualised Human Feedback Reveals About the Subjective and Multicultural Alignment of Large Language Models
• Compromesso! Italian Many-Shot Jailbreaks Undermine the Safety of Large Language Models
• My Answer is C: First-Token Probabilities Do Not Match Text Answers in Instruction-Tuned Language Models
• Political Compass or Spinning Arrow? Towards More Meaningful Evaluations for Values and Opinions in Large Language Models
• From Languages to Geographies: Towards Evaluating Cultural Bias in Hate Speech Datasets
• XSTest: A Test Suite for Identifying Exaggerated Safety Behaviors in Large Language Models
• Beyond Flesch-Kincaid: Prompt-based Metrics Improve Difficulty Classification of Educational Texts
• Safety-Tuned LLaMAs: Lessons From Improving the Safety of Large Language Models that Follow Instructions
• SafetyPrompts: a Systematic Review of Open Datasets for Evaluating and Improving Large Language Model Safety
• The Empty Signifier Problem: Towards Clearer Paradigms for Operationalising 'Alignment' in Large Language Models
• SimpleSafetyTests: a Test Suite for Identifying Critical Safety Risks in Large Language Models
• The Past, Present and Better Future of Feedback Learning in Large Language Models for Subjective Human Preferences and Values
• The Ecological Fallacy in Annotation: Modeling Human Label Variation goes beyond Sociodemographics
• Data-Efficient Strategies for Expanding Hate Speech Detection into Under-Resourced Languages
• Multilingual HateCheck: Functional Tests for Multilingual Hate Speech Detection Models
• Two Contrasting Data Annotation Paradigms for Subjective NLP Tasks