标签: evaluation - 66必读

涉及 OpenAI 模型的第三方网络评估

OpenAI explains recent third-party cybersecurity evaluation incidents and outlines new safeguards to strengthen AI model

evaluation and OpenAI

2026-08-05 OpenAI博客

智能AI None

必要但不充分：法学硕士法官安全评估中的温度控制和再现性

arXiv:2606.26185v1 Announce Type: new Abstract: LLM-as-judge ("grader") components are now standard in evaluation harnes

the grader temperature

2026-06-26 Hiroki Tamba

智能AI None

CELEUS：通过电子流程进行可认证且高效的法学硕士评估

20820v1 Announce Type: new Abstract: Can we trust evaluation scores to capture an LLM's true real-world performance。

the evaluation and

2026-06-23 Zhijian Zhou, Zesheng Ye, Zhaorun Chen, Bo Li, Feng Liu

智能AI None

人工智能代理可以综合科学结论吗？

11337v1 Announce Type: new Abstract: Scientific AI agents increasingly retrieve evidence, reason across sources, and syn

and that agents

2026-06-11 Hayoung Jung, Pedro Viana Diniz, Jos\'e Reinaldo Corr\^ea Roveda, Abner Fernandes da Silva, Haeun Jung, Enoch Tsai, Aleksandra Korolova, Manoel Horta Ribeiro

智能AI None

MacArena：在在线 macOS 环境上对计算机使用代理进行基准测试

arXiv:2606.06560v1 Announce Type: new Abstract: Computer-use agents (CUAs) operate graphical user interfaces (GUIs) thro

and macOS tasks

2026-06-08 Victor Muryn, Maksym Shamrai, Sofiia Mazepa, Yehor Khodysko

#evaluation

涉及 OpenAI 模型的第三方网络评估

必要但不充分：法学硕士法官安全评估中的温度控制和再现性

CELEUS：通过电子流程进行可认证且高效的法学硕士评估

人工智能代理可以综合科学结论吗？

MacArena：在在线 macOS 环境上对计算机使用代理进行基准测试