Adaptively Robust LLM Monitoring via Activation Watermarking
提出激活水印技术,使LLM监控对适应性攻击更鲁棒,防范离线搜索逃逸。
arXiv:2603.23171v3 Announce Type: replace-cross Abstract: Providers monitor deployed large language models (LLMs) to detect misuse that they cannot pr…
提出激活水印技术,使LLM监控对适应性攻击更鲁棒,防范离线搜索逃逸。
arXiv:2603.23171v3 Announce Type: replace-cross Abstract: Providers monitor deployed large language models (LLMs) to detect misuse that they cannot pr…
跨语言共情与文化适应性基准SPLIT,评估英/乌克兰语LLM的情感自然度与文化理解力。
arXiv:2607.02049v1 Announce Type: cross Abstract: Large Language Models are increasingly deployed in emotional-support contexts and crisis-related sit…
大模型能读懂文化语境,却在回应时忘得一干二净,这项研究揭示AI的文化适应性短板。
arXiv:2606.17688v1 Announce Type: new Abstract: Recent work has shown that LLMs overrepresent dominant cultures, particularly Western ones, while marg…
IT之家 6 月 10 日消息,据“海油工程微讯”今晚分享, 国内首台海上专用 PEM(质子交换膜)电解水制氢装置已于 6 月 5 日完成模拟海上浮式平台晃荡环境适应性试验 ,实验结果满足 17 级台风自存工况,为深远海新能源综合开发奠定坚实基础。 据介绍,PEM 制氢是一种先进的电解水制氢技术, …
LLM后训练新突破:AlphaToken通过路径感知机制对响应token进行估值,巧妙解耦适应性与稳定性,提升模型微调效率与鲁棒性。
arXiv:2606.01635v1 Announce Type: cross Abstract: Token selection is pivotal for effective LLM post-training. However, existing methods mostly rely on…
首个针对非洲语言的文化越狱基准,填补多语言安全评估空白
arXiv:2606.01322v1 Announce Type: cross Abstract: Safety evaluation of Large Language Models (LLMs) remains heavily English-centric, leaving Low-Resou…
神舟二十三号舷窗防护已做适应性改进,20天应急发射验证滚动备份策略有效性
IT之家 5 月 23 日消息,2026 年 5 月 23 日上午 9 时,神舟二十三号载人飞行任务新闻发布会在酒泉卫星发射中心召开,中国载人航天工程新闻发言人、工程办公室综合计划局局长张静波介绍了本次任务的相关情况。 去年底,中国载人航天工程实施了首次应急发射任务,备受国内外关注。对于经历了这次实…
针对LLM Agent在陌生环境中“过早利用”导致失败的问题,提出自主探索能力量化评估框架,助力构建更自适应的智能体。
arXiv:2605.16143v1 Announce Type: new Abstract: Large language model based agents often fail in unfamiliar environments due to premature exploitation:…