Datasets:

ARotting
/

snip-scope-exemplars

Tags:

mechanistic-interpretability

sparse-autoencoder

feature-discovery

License:

Dataset card Files Files and versions

snip-scope-exemplars / source

11 kB

1 contributor

History: 1 commit

ARotting's picture

Publish Held-out sparse-feature token exemplars

a80a585 verified 1 day ago