Aran Komatsuzaki· @arankomatsuzaki · X·· 6 天前AI 评分45
AI 导读
这个新基准用领域科学家开发的15个开放科学后续问题评估九个前沿智能体，产出135篇研究论文。 大多数论文未通过完整性检查，包括声称的工作与支撑性执行证据之间的不一致。
正文 · 原文
This new benchmark evaluates nine frontier agents on 15 open scientific follow-up problems developed with domain scientists, producing 135 research papers.
Most of the papers fail integrity checks, including discrepancies between claimed work and supporting execution evidence.
来源：Aran Komatsuzaki · x.com