Uncovering Branch specialization in InceptionV1 using k sparse autoencoders
Fuente:
arXiv
Saved in:
| Main Author: | |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866909580408127488 |
|---|---|
| author | Bozoukov, Matthew |
| author_facet | Bozoukov, Matthew |
| contents | Sparse Autoencoders (SAEs) have shown to find interpretable features in neural networks from polysemantic neurons caused by superposition. Previous work has shown SAEs are an effective tool to extract interpretable features from the early layers of InceptionV1. Since then, there have been many improvements to SAEs but branch specialization is still an enigma in the later layers of InceptionV1. We show various examples of branch specialization occuring in each layer of the mixed4a-4e branch, in the 5x5 branch and in one 1x1 branch. We also provide evidence to claim that branch specialization seems to be consistent across layers, similar features across the model will be localized in the same convolution size branches in their respective layer. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2504_11489 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Uncovering Branch specialization in InceptionV1 using k sparse autoencoders Bozoukov, Matthew Computer Vision and Pattern Recognition Sparse Autoencoders (SAEs) have shown to find interpretable features in neural networks from polysemantic neurons caused by superposition. Previous work has shown SAEs are an effective tool to extract interpretable features from the early layers of InceptionV1. Since then, there have been many improvements to SAEs but branch specialization is still an enigma in the later layers of InceptionV1. We show various examples of branch specialization occuring in each layer of the mixed4a-4e branch, in the 5x5 branch and in one 1x1 branch. We also provide evidence to claim that branch specialization seems to be consistent across layers, similar features across the model will be localized in the same convolution size branches in their respective layer. |
| title | Uncovering Branch specialization in InceptionV1 using k sparse autoencoders |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2504.11489 |