OpenAI· @OpenAI · X·· 2 小时前AI 评分44
AI 导读
@ApolloResearch 将评估我们的模型是否会暗中违背开发者的意图行事——以及我们的防护措施能否检测到这种行为。 它将审查来自训练、安全测试和内部使用的证据,以检验这种行为以及我们捕捉它的能力。
正文 · 原文
@ApolloResearch will assess whether our models could covertly work against their developers’ intentions—and whether our safeguards would detect it.
It will examine evidence from training, safety testing, and internal use to test for this behavior and our ability to catch it.
来源:OpenAI · x.com