1 paper · 1 filter
Jonathan Bennion, Shaona Ghosh, Mantek Singh +1
Various AI safety datasets have been developed to measure LLMs against evolving interpretations of harm. Our evaluation of five recently published open-source safety benchmarks rev…