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

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

Neuron 32 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: 0.9069. Min Range: -0.9069

Max Act: 0.9069. Min Act: -0.0001

Data Index: 1615327 (Python Code)

Max Activating Token Index: 463

Text #1

Max Range: 0.9069. Min Range: -0.9069

Max Act: 0.8842. Min Act: -0.0001

Data Index: 1423715 (Python Code)

Max Activating Token Index: 96

Text #2

Max Range: 0.9069. Min Range: -0.9069

Max Act: 0.7000. Min Act: -0.0001

Data Index: 947550 (C4 (Web Text))

Max Activating Token Index: 123

Text #3

Max Range: 0.9069. Min Range: -0.9069

Max Act: 0.7773. Min Act: -0.0001

Data Index: 497112 (C4 (Web Text))

Max Activating Token Index: 112

Text #4

Max Range: 0.9069. Min Range: -0.9069

Max Act: 0.6875. Min Act: -0.0001

Data Index: 1021297 (C4 (Web Text))

Max Activating Token Index: 244

Text #5

Max Range: 0.9069. Min Range: -0.9069

Max Act: 0.6744. Min Act: -0.0001

Data Index: 319798 (C4 (Web Text))

Max Activating Token Index: 25

Text #6

Max Range: 0.9069. Min Range: -0.9069

Max Act: 0.7469. Min Act: -0.0001

Data Index: 483348 (C4 (Web Text))

Max Activating Token Index: 884

Text #7

Max Range: 0.9069. Min Range: -0.9069

Max Act: 0.7058. Min Act: -0.0001

Data Index: 275707 (C4 (Web Text))

Max Activating Token Index: 86

Text #8

Max Range: 0.9069. Min Range: -0.9069

Max Act: 0.7260. Min Act: -0.0001

Data Index: 721596 (C4 (Web Text))

Max Activating Token Index: 22

Text #9

Max Range: 0.9069. Min Range: -0.9069

Max Act: 0.7671. Min Act: -0.0001

Data Index: 1460094 (Python Code)

Max Activating Token Index: 49

Text #10

Max Range: 0.9069. Min Range: -0.9069

Max Act: 0.7489. Min Act: -0.0001

Data Index: 1337636 (C4 (Web Text))

Max Activating Token Index: 14

Text #11

Max Range: 0.9069. Min Range: -0.9069

Max Act: 0.7086. Min Act: -0.0001

Data Index: 113536 (C4 (Web Text))

Max Activating Token Index: 29

Text #12

Max Range: 0.9069. Min Range: -0.9069

Max Act: 0.7210. Min Act: -0.0001

Data Index: 179100 (C4 (Web Text))

Max Activating Token Index: 17

Text #13

Max Range: 0.9069. Min Range: -0.9069

Max Act: 0.6798. Min Act: -0.0001

Data Index: 470234 (C4 (Web Text))

Max Activating Token Index: 95

Text #14

Max Range: 0.9069. Min Range: -0.9069

Max Act: 0.7127. Min Act: -0.0001

Data Index: 403859 (C4 (Web Text))

Max Activating Token Index: 18

Text #15

Max Range: 0.9069. Min Range: -0.9069

Max Act: 0.6801. Min Act: -0.0001

Data Index: 1189273 (C4 (Web Text))

Max Activating Token Index: 28

Text #16

Max Range: 0.9069. Min Range: -0.9069

Max Act: 0.7488. Min Act: -0.0001

Data Index: 92899 (C4 (Web Text))

Max Activating Token Index: 29

Text #17

Max Range: 0.9069. Min Range: -0.9069

Max Act: 0.7165. Min Act: -0.0001

Data Index: 32030 (C4 (Web Text))

Max Activating Token Index: 36

Text #18

Max Range: 0.9069. Min Range: -0.9069

Max Act: 0.6551. Min Act: -0.0001

Data Index: 745538 (C4 (Web Text))

Max Activating Token Index: 675

Text #19

Max Range: 0.9069. Min Range: -0.9069

Max Act: 0.6611. Min Act: -0.0001

Data Index: 393170 (C4 (Web Text))

Max Activating Token Index: 840