Fetching the paper…

ToolSword: Unveiling Safety Issues of Large Language Models in Tool Learning Across Three Stages · Around