Kimi K3.1 团队发现模型网络攻防能力随训练规模持续上升
Moonshot 在 Kimi K3.1 发布说明里自曝：训练数据里混入网络安全内容，模型漏洞挖掘能力随之变强，还讲了他们的移除实验。
Moonshot 在 Kimi K3.1 发布说明中披露，开发期间观察到漏洞研究性能随 post-training 规模提升而稳步变强。团队溯源发现，网络攻击相关内容混在通用 mid-training 语料中，相关任务也进入了 RL 环境 mix。内部模型的初步实验显示，若在 mid-training 或 RL 阶段都不做数据移除，下游 cyber 能力会大幅增加。这被视为大范围数据训练带来的自然副产品。
Models are are now approaching a massive general jump in cyber capabilities. From the Kimi K3.1 launch post:
'During development, we observed vulnerability research performance improving steadily as post-training scaled. Tracing the source of this trend, we found that cyber-related content was embedded within the general mid-training corpus, and cyber-related tasks were present in the RL environment mix — a natural byproduct of training on broad data. In preliminary experiments on internal models, we found that without data removal at either the mid-training or the RL stage, downstream cyber capability would increase substantially.'
