Science & math
Lean 4 disproof of the Köthe conjecture
In an Epoch AI eval, a pre-release GPT-6 Astra autonomously wrote a Lean 4 counterexample. Challenge/Solution, Comparator config, and CI are in the repo.
Back to projects
Science & math / GPT-6 Astra
已知 CVE 重发现评测
About this project
Aikido 提前获得模型；pass@3 为 29/32，pass@1 为 75%；3 次运行近 4000 美元；有提前访问关系；为已知漏洞基准，非本轮证明的新零日或生产审计
基准测试
Published prompt
The creator has not published a prompt for this project. Explore the original post or source code for more context.
Project media
Creator & source notes
Most entries have no play URL, source, or full prompt. Use the original post.