基础模型Foundation Model
Foundation Model
「大模型」的学术名称。指在海量数据上预训练、能适配一大堆任务的模型。The academic name for 'large models': pre-trained on massive data, adaptable to a wide range of tasks.
这个词由斯坦福 Bommasani 等人 2021 年提出。
指在大规模数据上预训练、可适配到广泛下游任务的模型。
作者用「基础」强调它们处于核心地位、但本身并不完整。
The term was proposed by Bommasani et al. at Stanford in 2021.
It refers to models pre-trained on broad data at scale and adaptable to a wide range of downstream tasks.
The authors chose 'foundation' to underscore their central yet incomplete character.
论文和监管文件常用「基础模型」而不是「大模型」。知道说的是同一类东西，就不会把两个词当成两件事。Papers and regulation say 'foundation model' where marketing says 'large model'. Knowing they name the same thing keeps you from treating them as two categories.
回到名词地图看第1站的其他 34 个词 →Back to the map: the other 34 terms in station 1 →
- Bommasani 等《On the Opportunities and Risks of Foundation Models》（Stanford HAI，2021）Bommasani et al., On the Opportunities and Risks of Foundation Models (Stanford HAI, 2021) ↗