netsec @lemmy.world repostbot33 @lemmy.world BOT 1y ago Researchers trying offensive capabilities of LLM agents arxiv.org /abs/2402.06664