Knowledge Neurons in Pretrained Transformers
arXiv:2104.08696
Abstract
Large-scale pretrained language models are surprisingly good at recalling factual knowledge presented in the training corpus. In this paper, we present preliminary studies on how factual knowledge is stored in pretrained Transformers by introducing the concept of knowledge neurons. Specifically, we examine the fill-in-the-blank cloze task for BERT. Given a relational fact, we propose a knowledge attribution method to identify the neurons that express the fact. We find that the activation of such knowledge neurons is positively correlated to the expression of their corresponding facts. In our case studies, we attempt to leverage knowledge neurons to edit (such as update, and erase) specific factual knowledge without fine-tuning. Our results shed light on understanding the storage of knowledge within pretrained Transformers. The code is available at https://github.com/Hunter-DDM/knowledge-neurons.
arXiv commentACL 2022
Cited by in corpus (5)
- Hybrid Transformer with Multi-level Fusion for Multimodal Knowledge Graph Completion
- On Transferability of Prompt Tuning for Natural Language Processing
- A Survey on Symbolic Knowledge Distillation of Large Language Models
- Distilling Relation Embeddings from Pre-trained Language Models
- Language Anisotropic Cross-Lingual Model Editing