BERT Busters: Outlier Dimensions that Disrupt Transformers
Multiple studies have shown that Transformers are remarkably robust to pruning. Contrary to this received wisdom, we demonstrate that pre-trained Transformer encoders are surprisingly fragile to the removal of a very small number of features in the layer outputs (<0.0001% of model weights). In case of BERT and other pre-trained encoder Transformers, the affected component is the scaling factors and biases in the LayerNorm. The outliers are high-magnitude normalization parameters that emerge early in pre-training and show up consistently in the same dimensional position throughout the model. We show that disabling them significantly degrades both the MLM loss and the downstream task performance. This effect is observed across several BERT-family models and other popular pre-trained Transformer architectures, including BART, XLNet and ELECTRA; we also show a similar effect in GPT-2.
Code (0)
등록된 구현이 없습니다.
Tasks
PositionMethods 이 논문이 사용한 방법론
Similar Papers 제목 키워드 기반
Outliers Dimensions that Disrupt Transformers Are Driven by Frequency
While Transformer-based language models are generally very robust to pruning, there is the recently discovered outlier phenomenon: disabling only 48 out of 110M parameters in BERT-base drops its performance by nearly 30%…
IITR CodeBusters at SemEval-2022 Task 5: Misogyny Identification using Transformers
This paper presents our submission to task 5 ( Multimedia Automatic Misogyny Identification) of the SemEval 2022 competition. The purpose of the task is to identify given memes as misogynistic or not and further label th…
NCU-NLP at ROCLING-2021 Shared Task: Using MacBERT Transformers for Dimensional Sentiment Analysis
We use the MacBERT transformers and fine-tune them to ROCLING-2021 shared tasks using the CVAT and CVAS data. We compare the performance of MacBERT with the other two transformers BERT and RoBERTa in the valence and arou…
Sentiment AnalysisAn Isotropy Analysis in the Multilingual BERT Embedding Space
Several studies have explored various advantages of multilingual pre-trained models (such as multilingual BERT) in capturing shared linguistic knowledge. However, less attention has been paid to their limitations. In thi…
Semantic SimilaritySemantic Textual SimilarityExploring Anisotropy and Outliers in Multilingual Language Models for Cross-Lingual Semantic Sentence Similarity
Previous work has shown that the representations output by contextual language models are more anisotropic than static type embeddings, and typically display outlier dimensions. This seems to be true for both monolingual…
Semantic SimilaritySemantic Textual SimilaritySentenceSentence Similarity