Model: SoLU Model: 10 Layers, 5120 Neurons per Layer

Dataset: 80% C4 (Web Text) and 20% Python Code

Neuron 1257 in Layer 7

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-10l')

Text #0

Max Range: 0.6261. Min Range: -0.6261

Max Act: 0.6122. Min Act: -0.0000

Data Index: 2730 (C4 (Web Text))

Max Activating Token Index: 966

Text #1

Max Range: 0.6261. Min Range: -0.6261

Max Act: 0.4386. Min Act: -0.0000

Data Index: 1221553 (C4 (Web Text))

Max Activating Token Index: 285

Text #2

Max Range: 0.6261. Min Range: -0.6261

Max Act: 0.6148. Min Act: -0.0000

Data Index: 747927 (C4 (Web Text))

Max Activating Token Index: 87

Text #3

Max Range: 0.6261. Min Range: -0.6261

Max Act: 0.6261. Min Act: -0.0000

Data Index: 200448 (C4 (Web Text))

Max Activating Token Index: 600

Text #4

Max Range: 0.6261. Min Range: -0.6261

Max Act: 0.4826. Min Act: -0.0000

Data Index: 767878 (C4 (Web Text))

Max Activating Token Index: 943

Text #5

Max Range: 0.6261. Min Range: -0.6261

Max Act: 0.4755. Min Act: -0.0000

Data Index: 504065 (C4 (Web Text))

Max Activating Token Index: 924

Text #6

Max Range: 0.6261. Min Range: -0.6261

Max Act: 0.4378. Min Act: -0.0000

Data Index: 16703 (C4 (Web Text))

Max Activating Token Index: 608

Text #7

Max Range: 0.6261. Min Range: -0.6261

Max Act: 0.4646. Min Act: -0.0000

Data Index: 242350 (C4 (Web Text))

Max Activating Token Index: 37

Text #8

Max Range: 0.6261. Min Range: -0.6261

Max Act: 0.4424. Min Act: -0.0000

Data Index: 526878 (C4 (Web Text))

Max Activating Token Index: 224

Text #9

Max Range: 0.6261. Min Range: -0.6261

Max Act: 0.4220. Min Act: -0.0000

Data Index: 990323 (C4 (Web Text))

Max Activating Token Index: 452

Text #10

Max Range: 0.6261. Min Range: -0.6261

Max Act: 0.4516. Min Act: -0.0000

Data Index: 1491031 (Python Code)

Max Activating Token Index: 920

Text #11

Max Range: 0.6261. Min Range: -0.6261

Max Act: 0.4342. Min Act: -0.0000

Data Index: 837448 (C4 (Web Text))

Max Activating Token Index: 561

Text #12

Max Range: 0.6261. Min Range: -0.6261

Max Act: 0.4238. Min Act: -0.0000

Data Index: 1436541 (Python Code)

Max Activating Token Index: 356

Text #13

Max Range: 0.6261. Min Range: -0.6261

Max Act: 0.4176. Min Act: -0.0000

Data Index: 1338675 (C4 (Web Text))

Max Activating Token Index: 281

Text #14

Max Range: 0.6261. Min Range: -0.6261

Max Act: 0.4295. Min Act: -0.0000

Data Index: 163810 (C4 (Web Text))

Max Activating Token Index: 824

Text #15

Max Range: 0.6261. Min Range: -0.6261

Max Act: 0.4149. Min Act: -0.0000

Data Index: 941396 (C4 (Web Text))

Max Activating Token Index: 419

Text #16

Max Range: 0.6261. Min Range: -0.6261

Max Act: 0.4004. Min Act: -0.0000

Data Index: 1365227 (Python Code)

Max Activating Token Index: 190

Text #17

Max Range: 0.6261. Min Range: -0.6261

Max Act: 0.3993. Min Act: -0.0000

Data Index: 1090051 (C4 (Web Text))

Max Activating Token Index: 430

Text #18

Max Range: 0.6261. Min Range: -0.6261

Max Act: 0.4271. Min Act: -0.0000

Data Index: 166981 (C4 (Web Text))

Max Activating Token Index: 966

Text #19

Max Range: 0.6261. Min Range: -0.6261

Max Act: 0.4330. Min Act: -0.0000

Data Index: 1371785 (Python Code)

Max Activating Token Index: 818