你的大模型安全吗？360大模型卫士检测系统，给AI做个全面“体检”
"模型风险" 相关文章
Developers beware: Google’s Gemma model controversy exposes model lifecycle risks
Anthropic's Pilot Sabotage Risk Report
Can we steer AI models toward safer actions by making these instrumentally useful?
’Getting the Models Right’: How to Value Hard-to-Price Assets
We are likely in an AI overhang, and this is bad.
Jailbreak迎来“最后一卷”？港科大用“内容评分”重塑大模型越狱评估范式
黑化威胁操纵人类，Claude勒索，o1自主逃逸，人类「执剑人」紧急上线
Contrived evaluations are useful evaluations
Agentic Misalignment: How LLMs Could be Insider Threats
如果竞争对手发布“高风险”AI OpenAI 可能会“调整”其安全措施
38.8 - David Duvenaud on Sabotage Evaluations and the Post-AGI Future
[国 际] 合成数据能否让AI模型精确可靠？
Distinguish worst-case analysis from instrumental training-gaming
Has Eliezer publicly and satisfactorily responded to attempted rebuttals of the analogy to evolution?
