Measuring LLMs' ability to develop exploitsred.anthropic.com5 points·Kneenex··0 commentsOpen articleSaveView on HN