search

DavidUdell / sparse_circuit_discovery

Circuit discovery in GPT-2 small, using sparse autoencoding
MIT License
7 stars 1 forks source link

Layer indexed DIMS_PLOTTED_LIST #28

Closed DavidUdell closed 10 months ago

DavidUdell commented 10 months ago

Generally, there won't exist the same meaningful indices across layer autoencoders. The syntax I use to manually set a DIMS_LIST should reflect this.