Searching for "Harmful Refusal": A Psychometric Audit of an AI Safety Benchmark

arXiv:2610.12409 · cs.AI · Submitted 2026-10-08 · Read on arXiv

cs.AI

Submitted: 2026-10-08

Updated: 2026-10-08

Comments: COLM 2026, AI for Measurement Science (AIMS) Workshop

Code: https://github.com/cmstewart/harmful-refusal-audit

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Terminology

• Benchmark — 1× in this paper · explained in 2 episodes

Sources

• Concrete Problems in AI Safety
• Red Teaming Language Models to Reduce Harms: Methods, Scaling Behaviors, and Lessons Learned
• Unsolved Problems in ML Safety
• SimpleSafetyTests: a Test Suite for Identifying Critical Safety Risks in Large Language Models

Related papers

• MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
• Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
• Agents' Last Exam
• The Clinician's Veto: Navigating Trust, Liability, and Uncertainty in Autonomous AI Prescribing
• MindHelper: Closed-Loop Embodied Mental-State Reasoning for Precision Intervention
• Incumbent Advantage: Brand Bias and Cognitive Manipulation Dynamics in LLM Recommendation Systems