arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2106.08367cs.CL
收藏

Transformer语言模型可以使用哪些上下文特征?

What Context Features Can Transformer Language Models Use?

Massachusetts Institute of Technology(麻省理工学院)

Joe O'Connor, Jacob Andreas

摘要:

基于transformer的语言模型受益于对数百到数千个先前标记的上下文的调节。这些上下文的哪些方面有助于准确的模型预测?我们描述了一系列的实验,测量可用的信息,有选择地消融词汇和结构信息的Transformer语言模型上训练的英语维基百科。在中、长范围的语境中,我们发现一些极具破坏性的语境操作--包括打乱句子中的词序和删除除名词以外的所有单词--只删除了不到15%的可用信息。我们的研究结果表明,长的上下文,但不是他们详细的句法和命题内容,是重要的低困惑电流Transformer语言模型。

英文摘要:

Transformer-based language models benefit from conditioning on contexts of hundreds to thousands of previous tokens. What aspects of these contexts contribute to accurate model prediction? We describe a series of experiments that measure usable information by selectively ablating lexical and structural information in transformer language models trained on English Wikipedia. In both mid- and long-range contexts, we find that several extremely destructive context manipulations -- including shuffling word order within sentences and deleting all words other than nouns -- remove less than 15% of the usable information. Our results suggest that long contexts, but not their detailed syntactic and propositional content, are important for the low perplexity of current transformer language models.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑