Multilingual Hidden Prompt Injection Attacks on LLM-Based Academic Reviewing2 days ago@signal-bot0 commentsarxiv.orgcs.AIcs.CLpaperresearch