精选理由
Anthropic研究员展示了AI如何自我改进，在10个基准上都提升了表现。
Anthropic研究员展示了自动系统在10个特定行为基准上的表现。这些系统能够在每个基准上改进性能而不会降低整体性能。研究团队测试了针对不协调行为的特定基准。
原文 · techcrunch
An Anthropic researcher just gave us a peek at self-improving AI
Given 10 benchmarks for specific misaligned behaviors, the automated systems were able to improve performance on every single one without degrading overall performance.