Estimating the difficulty of a dataset typically involves comparing state-of-the-art models to
humans; the bigger the performance gap, the harder the dataset is said to be. However, this …

In the last half-decade, the field of natural language processing (NLP) has undergone two
major transitions: the switch to neural networks as the primary modeling paradigm and the …

Individual neurons participate in the representation of multiple high-level concepts. To what
extent can different interpretability methods successfully disentangle these roles? To help …

Multi-step reasoning ability is fundamental to many natural language tasks, yet it is unclear
what constitutes a good reasoning chain and how to evaluate them. Most existing methods …

Given the prevalence of pre-trained contextualized representations in today's NLP, there
have been many efforts to understand what information they contain, and why they seem to …

A central quest of probing is to uncover how pre-trained models encode a linguistic property
within their representations. An encoding, however, might be spurious-ie, the model might …

In this paper, we investigate to which extent contextual neural language models (LMs)
implicitly learn syntactic structure. More concretely, we focus on constituent structure as …

摘要随着大规模预训练模型的广泛应用, 自然语言处理的多个领域(如文本分类和机器翻译)
取得了长足的发展. 然而, 受限于预训练模型的“黑盒” 特性, 其内部的决策模式以及编码的知识 …

Because meaning can often be inferred from lexical semantics alone, word order is often a
redundant cue in natural language. For example, the words chopped, chef, and onion are …

Understanding which concepts models can and cannot represent has been fundamental to
many tasks: from effective and responsible use of models to detecting out of distribution …