Abstract
Understanding robustness is essential for building reliable NLP systems. Unfortunately, in the context of machine translation, previous work mainly focused on documenting robustness failures or improving robustness. In contrast, we study robustness from a model representation perspective by looking at internal model representations of ungrammatical inputs and how they evolve through model layers. For this purpose, we perform Grammatical Error Detection (GED) probing and representational similarity analysis. Our findings indicate that the encoder first detects the grammatical error, then corrects it by moving its representation toward the correct form. To understand what contributes to this process, we turn to the attention mechanism where we identify what we term Robustness Heads. We find that Robustness Heads attend to interpretable linguistic units when responding to grammatical errors, and that when we fine-tune models for robustness, they tend to rely more on Robustness Heads for updating the ungrammatical word representation.
Abstract (translated)
理解稳健性对于构建可靠的自然语言处理系统至关重要。不幸的是，在机器翻译的背景下，之前的工作主要集中在记录稳健性的失败案例或改进稳健性上。相比之下，我们从模型表示的角度研究稳健性，通过观察非语法输入在各个模型层中的内部表示以及它们如何演变来实现这一目标。为此，我们进行了语法错误检测（GED）探测和表征相似度分析。我们的发现表明，编码器首先检测到语法错误，然后通过将表示形式移动到正确的形式来进行修正。为了了解推动这一过程的因素，我们转向注意力机制，在那里我们识别出了我们称之为“稳健性头”的部分。我们发现，“稳健性头”在响应语法错误时会关注可解释的语言单元，并且当我们针对稳健性进行模型微调时，它们倾向于更多地依赖于“稳健性头”来更新非语法单词的表示形式。
URL
https://arxiv.org/abs/2505.21224
