引用og_kalu · 2023-05
“A world model is so obvious papers like these are more confirmation than surprise”
称世界模型显而易见,这类论文更多是印证而非惊喜
探针实验显示代码语言模型仅靠预测下一词就学到程序语义
论文给出证据:仅以下一词预测训练的代码语言模型,也能学会表示程序的形式语义。作者在一种控制二维网格世界导航的领域特定语言的合成程序语料上训练 Transformer,每个程序前附有若干输入输出网格状态作为(部分)规范。探针分类器能从隐藏状态中提取出越来越准确的、未被观察到的中间网格状态,表明模型涌现出形式意义上解释程序的能力;作者还设计了干预式基线来区分表示究竟来自语言模型还是探针本身。
在争论语言模型是否只是统计鹦鹉时,评论者把它当作 LM 内部形成程序语义或世界模型表示的证据。
“A world model is so obvious papers like these are more confirmation than surprise”
“3 of them don't even have anything to do with a existing dataset testing. so yeah, actual.”
“Another wrong statement, you're on a roll today.”
“Case in point:”
“We have other studies indicating that there appear to be internal representations of concepts: Emergent Representations of Program Semantics in Language Models Trained on Programs”
“LLMs build models of the world[2].”