Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2401.16731 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917577875259392 |
|---|---|
| author | Mondal, Shrayani Garodia, Rishabh Qureshi, Arbaaz Lee, Taesung Park, Youngja |
| author_facet | Mondal, Shrayani Garodia, Rishabh Qureshi, Arbaaz Lee, Taesung Park, Youngja |
| contents | Recent developments in transformer-based language models have allowed them to capture a wide variety of world knowledge that can be adapted to downstream tasks with limited resources. However, what pieces of information are understood in these models is unclear, and neuron-level contributions in identifying them are largely unknown. Conventional approaches in neuron explainability either depend on a finite set of pre-defined descriptors or require manual annotations for training a secondary model that can then explain the neurons of the primary model. In this paper, we take BERT as an example and we try to remove these constraints and propose a novel and scalable framework that ties textual descriptions to neurons. We leverage the potential of generative language models to discover human-interpretable descriptors present in a dataset and use an unsupervised approach to explain neurons with these descriptors. Through various qualitative and quantitative analyses, we demonstrate the effectiveness of this framework in generating useful data-specific descriptors with little human involvement in identifying the neurons that encode these descriptors. In particular, our experiment shows that the proposed approach achieves 75% precision@2, and 50% recall@2 |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2401_16731 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Towards Generating Informative Textual Description for Neurons in Language Models Mondal, Shrayani Garodia, Rishabh Qureshi, Arbaaz Lee, Taesung Park, Youngja Computation and Language Artificial Intelligence Recent developments in transformer-based language models have allowed them to capture a wide variety of world knowledge that can be adapted to downstream tasks with limited resources. However, what pieces of information are understood in these models is unclear, and neuron-level contributions in identifying them are largely unknown. Conventional approaches in neuron explainability either depend on a finite set of pre-defined descriptors or require manual annotations for training a secondary model that can then explain the neurons of the primary model. In this paper, we take BERT as an example and we try to remove these constraints and propose a novel and scalable framework that ties textual descriptions to neurons. We leverage the potential of generative language models to discover human-interpretable descriptors present in a dataset and use an unsupervised approach to explain neurons with these descriptors. Through various qualitative and quantitative analyses, we demonstrate the effectiveness of this framework in generating useful data-specific descriptors with little human involvement in identifying the neurons that encode these descriptors. In particular, our experiment shows that the proposed approach achieves 75% precision@2, and 50% recall@2 |
| title | Towards Generating Informative Textual Description for Neurons in Language Models |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2401.16731 |