人机协作写作竟藏越狱风险？新基准揭示大模型安全新盲区
arXiv:2604.19274v2 Announce Type: replace Abstract: Large language models (LLMs) are increasingly used as co-authors in collaborative writing, where u…
AI伴侣暗藏风险？新框架评估大模型社交对话中的负面动态。
arXiv:2605.30654v1 Announce Type: cross Abstract: Large language models (LLMs) are increasingly used as conversational partners for companionship, emo…
400次重复实验揭示：大模型做黑客竟如此「不稳定」？首个LLM渗透测试一致性量化研究。
arXiv:2605.30096v1 Announce Type: cross Abstract: Large language models (LLMs) can autonomously conduct multi-stage cyber attacks, but the consistency…
最新研究揭示指令微调LLM在长上下文输入中，对有害句子的敏感性存在显著风险。
arXiv:2510.05864v2 Announce Type: replace Abstract: Large language models (LLMs) increasingly operate on long inputs, yet their behavior when harmful …
基于LLM隐藏状态探测的多语言幻觉检测新方法，有效应对非英语场景的可靠性挑战。
arXiv:2605.24919v1 Announce Type: new Abstract: Hallucinations in Large Language Models (LLMs) represent a critical barrier to their reliable deployme…