XSTest

A new paper, 'Safety for Whom? Boundary-Aware Self-Distillation for Controlled LLM Safety Refusal,' argues that topic-level safety taxonomies like LlamaGuard-3 are too blunt for real deployments, whic…

A new arXiv study (2609.04482v1) introduces a boundary-aware self-distillation framework for controlled LLM safety refusal, showing that escalating retries reduce prompts without accepted refusal trac…

A new arXiv preprint (2608.02665v1) finds that evaluating large language models on a single canonical prompt surface underestimates unsafe compliance by 3.3 to 12.9 percentage points across five model…