Tag
This paper explores interpretation difficulty in harmful Discord chats related to cybercrime, evaluating human and LLM interpretations under different context conditions and proposing a classification of factors that make such chats hard to interpret.