Saved in:
| Main Authors: | , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2404.15255 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866910419587694592 |
|---|---|
| author | Heimersheim, Stefan Nanda, Neel |
| author_facet | Heimersheim, Stefan Nanda, Neel |
| contents | Activation patching is a popular mechanistic interpretability technique, but has many subtleties regarding how it is applied and how one may interpret the results. We provide a summary of advice and best practices, based on our experience using this technique in practice. We include an overview of the different ways to apply activation patching and a discussion on how to interpret the results. We focus on what evidence patching experiments provide about circuits, and on the choice of metric and associated pitfalls. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2404_15255 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | How to use and interpret activation patching Heimersheim, Stefan Nanda, Neel Machine Learning Activation patching is a popular mechanistic interpretability technique, but has many subtleties regarding how it is applied and how one may interpret the results. We provide a summary of advice and best practices, based on our experience using this technique in practice. We include an overview of the different ways to apply activation patching and a discussion on how to interpret the results. We focus on what evidence patching experiments provide about circuits, and on the choice of metric and associated pitfalls. |
| title | How to use and interpret activation patching |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2404.15255 |