Super-NaturalInstructions: Generalization via Declarative Instructions on 1600+ NLP Tasks
超自然指令：通过声明式指令在 1600 多个 NLP 任务上进行泛化.
Rethinking the Role of Demonstrations: What Makes In-Context Learning Work?
重新思考演示的作用：什么导致上下文学习有效？
40篇短文搞定3500个单词
Learning 3,500 words from 40 short articles.
琴生不等式(Jenson’s Inequality)
Jenson’s Inequality.
Emergent Abilities of Large Language Models
大型语言模型的涌现能力.
Beyond the Imitation Game: Quantifying and extrapolating the capabilities of language models
超越模仿游戏：量化和推断语言模型的能力.