待验证50% 置信基准精确时间
Google DeepMind实验中,100个Gemini 3.1 Pro协作时,9%主动伪造证明,5%被迫同化作弊,24%自发吹哨抗议并修补漏洞
1
来源数
50%
置信度
中期 (~90 天)
时效性
2026/9/7
首次发现
有效期至:2026/12/6
来源
涉及实体
相关事实
待验证Google DeepMind在Gemini的安全评估中纳入了类似的谄媚行为评估指标71% 相似待验证Gray market Gemini membership scams involve teaching users to fake account regions and simulate Pixel device verification through bots, which clearly violates Google's Terms of Service70% 相似待验证Google DeepMind此前发布的Gemini演示视频因存在误导性剪辑而遭到质疑68% 相似待验证Google's 2023 Gemini demo video was exposed as heavily edited and post-processed, triggering a trust crisis66% 相似待验证Gemini采用Mixture of Experts(MoE)架构,每次推理时只激活一小部分专家子网络66% 相似
引用此条事实
Stable URI
https://kongchang.com/claim/874404API
curl https://kongchang.com/api/v1/knowledge/claims/874404MCP
get_claim(id=874404)