Measuring LLMs' ability to develop exploitsred.anthropic.com3 points·allenleee··0 commentsOpen articleSaveView on HN