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

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

Neuron 1603 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.9328. Min Range: -0.9328

Max Act: 0.8981. Min Act: -0.0001

Data Index: 550230 (C4 (Web Text))

Max Activating Token Index: 172

Text #1

Max Range: 0.9328. Min Range: -0.9328

Max Act: 0.9328. Min Act: -0.0001

Data Index: 1057253 (C4 (Web Text))

Max Activating Token Index: 774

Text #2

Max Range: 0.9328. Min Range: -0.9328

Max Act: 0.8458. Min Act: -0.0001

Data Index: 3296 (C4 (Web Text))

Max Activating Token Index: 707

Text #3

Max Range: 0.9328. Min Range: -0.9328

Max Act: 0.8383. Min Act: -0.0001

Data Index: 1107449 (C4 (Web Text))

Max Activating Token Index: 782

Text #4

Max Range: 0.9328. Min Range: -0.9328

Max Act: 0.8209. Min Act: -0.0001

Data Index: 62783 (C4 (Web Text))

Max Activating Token Index: 34

Text #5

Max Range: 0.9328. Min Range: -0.9328

Max Act: 0.8647. Min Act: -0.0001

Data Index: 239947 (C4 (Web Text))

Max Activating Token Index: 22

Text #6

Max Range: 0.9328. Min Range: -0.9328

Max Act: 0.8411. Min Act: -0.0001

Data Index: 341611 (C4 (Web Text))

Max Activating Token Index: 40

Text #7

Max Range: 0.9328. Min Range: -0.9328

Max Act: 0.7780. Min Act: -0.0001

Data Index: 1108205 (C4 (Web Text))

Max Activating Token Index: 57

Text #8

Max Range: 0.9328. Min Range: -0.9328

Max Act: 0.8028. Min Act: -0.0001

Data Index: 835063 (C4 (Web Text))

Max Activating Token Index: 34

Text #9

Max Range: 0.9328. Min Range: -0.9328

Max Act: 0.9234. Min Act: -0.0001

Data Index: 740089 (C4 (Web Text))

Max Activating Token Index: 218

Text #10

Max Range: 0.9328. Min Range: -0.9328

Max Act: 0.8069. Min Act: -0.0001

Data Index: 1008839 (C4 (Web Text))

Max Activating Token Index: 37

Text #11

Max Range: 0.9328. Min Range: -0.9328

Max Act: 0.7895. Min Act: -0.0001

Data Index: 894336 (C4 (Web Text))

Max Activating Token Index: 42

Text #12

Max Range: 0.9328. Min Range: -0.9328

Max Act: 0.8122. Min Act: -0.0001

Data Index: 110405 (C4 (Web Text))

Max Activating Token Index: 39

Text #13

Max Range: 0.9328. Min Range: -0.9328

Max Act: 0.8278. Min Act: -0.0001

Data Index: 1025511 (C4 (Web Text))

Max Activating Token Index: 40

Text #14

Max Range: 0.9328. Min Range: -0.9328

Max Act: 0.8264. Min Act: -0.0001

Data Index: 485996 (C4 (Web Text))

Max Activating Token Index: 35

Text #15

Max Range: 0.9328. Min Range: -0.9328

Max Act: 0.8276. Min Act: -0.0001

Data Index: 984225 (C4 (Web Text))

Max Activating Token Index: 853

Text #16

Max Range: 0.9328. Min Range: -0.9328

Max Act: 0.7895. Min Act: -0.0001

Data Index: 319985 (C4 (Web Text))

Max Activating Token Index: 532

Text #17

Max Range: 0.9328. Min Range: -0.9328

Max Act: 0.7867. Min Act: -0.0001

Data Index: 1254146 (C4 (Web Text))

Max Activating Token Index: 56

Text #18

Max Range: 0.9328. Min Range: -0.9328

Max Act: 0.7830. Min Act: -0.0001

Data Index: 468081 (C4 (Web Text))

Max Activating Token Index: 48

Text #19

Max Range: 0.9328. Min Range: -0.9328

Max Act: 0.7374. Min Act: -0.0001

Data Index: 535928 (C4 (Web Text))

Max Activating Token Index: 20