The Decoder· Jonathan Kemper·· 4 小时前AI 评分68
Sakana AI 提出 MASS 方法实现语言模型递归自改进但存在共享盲区
Sakana AI's self-improving model gets better at solving and judging, but can't catch its own blind spots
AI 导读
Sakana AI 与 UC Berkeley 研究人员提出“多智能体自我监督”(MASS)方法,使单一语言模型通过解决任务、评估结果和优化工作流实现递归自改进。
来源:The Decoder · the-decoder.com