Model: SoLU Model: 8 Layers, 4096 Neurons per Layer

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

Neuron 3872 in Layer 6

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-8l')

Text #0

Max Range: 0.7606. Min Range: -0.7606

Max Act: 0.7606. Min Act: -0.0001

Data Index: 1341034 (C4 (Web Text))

Max Activating Token Index: 340

Text #1

Max Range: 0.7606. Min Range: -0.7606

Max Act: 0.7458. Min Act: -0.0000

Data Index: 727637 (C4 (Web Text))

Max Activating Token Index: 724

Text #2

Max Range: 0.7606. Min Range: -0.7606

Max Act: 0.6754. Min Act: -0.0000

Data Index: 675628 (C4 (Web Text))

Max Activating Token Index: 61

Text #3

Max Range: 0.7606. Min Range: -0.7606

Max Act: 0.6719. Min Act: -0.0000

Data Index: 1501237 (Python Code)

Max Activating Token Index: 852

Text #4

Max Range: 0.7606. Min Range: -0.7606

Max Act: 0.6263. Min Act: -0.0000

Data Index: 952426 (C4 (Web Text))

Max Activating Token Index: 760

Text #5

Max Range: 0.7606. Min Range: -0.7606

Max Act: 0.6118. Min Act: -0.0000

Data Index: 404989 (C4 (Web Text))

Max Activating Token Index: 880

Text #6

Max Range: 0.7606. Min Range: -0.7606

Max Act: 0.6694. Min Act: -0.0000

Data Index: 870442 (C4 (Web Text))

Max Activating Token Index: 279

Text #7

Max Range: 0.7606. Min Range: -0.7606

Max Act: 0.6212. Min Act: -0.0000

Data Index: 1124396 (C4 (Web Text))

Max Activating Token Index: 687

Text #8

Max Range: 0.7606. Min Range: -0.7606

Max Act: 0.6102. Min Act: -0.0000

Data Index: 45072 (C4 (Web Text))

Max Activating Token Index: 64

Text #9

Max Range: 0.7606. Min Range: -0.7606

Max Act: 0.5983. Min Act: -0.0001

Data Index: 1239808 (C4 (Web Text))

Max Activating Token Index: 804

Text #10

Max Range: 0.7606. Min Range: -0.7606

Max Act: 0.5914. Min Act: -0.0000

Data Index: 351223 (C4 (Web Text))

Max Activating Token Index: 877

Text #11

Max Range: 0.7606. Min Range: -0.7606

Max Act: 0.5666. Min Act: -0.0000

Data Index: 91360 (C4 (Web Text))

Max Activating Token Index: 558

Text #12

Max Range: 0.7606. Min Range: -0.7606

Max Act: 0.6137. Min Act: -0.0000

Data Index: 130964 (C4 (Web Text))

Max Activating Token Index: 665

Text #13

Max Range: 0.7606. Min Range: -0.7606

Max Act: 0.5652. Min Act: -0.0000

Data Index: 1298475 (C4 (Web Text))

Max Activating Token Index: 928

Text #14

Max Range: 0.7606. Min Range: -0.7606

Max Act: 0.5685. Min Act: -0.0000

Data Index: 877163 (C4 (Web Text))

Max Activating Token Index: 417

Text #15

Max Range: 0.7606. Min Range: -0.7606

Max Act: 0.5145. Min Act: -0.0001

Data Index: 1621556 (Python Code)

Max Activating Token Index: 936

Text #16

Max Range: 0.7606. Min Range: -0.7606

Max Act: 0.5561. Min Act: -0.0001

Data Index: 740937 (C4 (Web Text))

Max Activating Token Index: 9

Text #17

Max Range: 0.7606. Min Range: -0.7606

Max Act: 0.5731. Min Act: -0.0000

Data Index: 1113067 (C4 (Web Text))

Max Activating Token Index: 472

Text #18

Max Range: 0.7606. Min Range: -0.7606

Max Act: 0.5532. Min Act: -0.0000

Data Index: 593904 (C4 (Web Text))

Max Activating Token Index: 566

Text #19

Max Range: 0.7606. Min Range: -0.7606

Max Act: 0.5464. Min Act: -0.0000

Data Index: 193654 (C4 (Web Text))

Max Activating Token Index: 21