Model: SoLU Model: 12 Layers, 6144 Neurons per Layer

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

Neuron 4770 in Layer 8

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-12l')

Text #0

Max Range: 0.1045. Min Range: -0.1045

Max Act: 0.1036. Min Act: -0.0000

Data Index: 759948 (C4 (Web Text))

Max Activating Token Index: 261

Text #1

Max Range: 0.1045. Min Range: -0.1045

Max Act: 0.0999. Min Act: -0.0000

Data Index: 1571489 (Python Code)

Max Activating Token Index: 187

Text #2

Max Range: 0.1045. Min Range: -0.1045

Max Act: 0.0951. Min Act: -0.0000

Data Index: 837580 (C4 (Web Text))

Max Activating Token Index: 186

Text #3

Max Range: 0.1045. Min Range: -0.1045

Max Act: 0.0940. Min Act: -0.0000

Data Index: 1442759 (Python Code)

Max Activating Token Index: 962

Text #4

Max Range: 0.1045. Min Range: -0.1045

Max Act: 0.0940. Min Act: -0.0000

Data Index: 482055 (C4 (Web Text))

Max Activating Token Index: 516

Text #5

Max Range: 0.1045. Min Range: -0.1045

Max Act: 0.0962. Min Act: -0.0000

Data Index: 243284 (C4 (Web Text))

Max Activating Token Index: 162

Text #6

Max Range: 0.1045. Min Range: -0.1045

Max Act: 0.0917. Min Act: -0.0000

Data Index: 1207318 (C4 (Web Text))

Max Activating Token Index: 169

Text #7

Max Range: 0.1045. Min Range: -0.1045

Max Act: 0.0975. Min Act: -0.0000

Data Index: 1500054 (Python Code)

Max Activating Token Index: 684

Text #8

Max Range: 0.1045. Min Range: -0.1045

Max Act: 0.0909. Min Act: -0.0000

Data Index: 1486851 (Python Code)

Max Activating Token Index: 480

Text #9

Max Range: 0.1045. Min Range: -0.1045

Max Act: 0.1045. Min Act: -0.0000

Data Index: 1426363 (Python Code)

Max Activating Token Index: 355

Text #10

Max Range: 0.1045. Min Range: -0.1045

Max Act: 0.0867. Min Act: -0.0000

Data Index: 319433 (C4 (Web Text))

Max Activating Token Index: 197

Text #11

Max Range: 0.1045. Min Range: -0.1045

Max Act: 0.0927. Min Act: -0.0000

Data Index: 1364084 (Python Code)

Max Activating Token Index: 119

Text #12

Max Range: 0.1045. Min Range: -0.1045

Max Act: 0.0947. Min Act: -0.0000

Data Index: 538852 (C4 (Web Text))

Max Activating Token Index: 105

Text #13

Max Range: 0.1045. Min Range: -0.1045

Max Act: 0.0934. Min Act: -0.0000

Data Index: 1442437 (Python Code)

Max Activating Token Index: 347

Text #14

Max Range: 0.1045. Min Range: -0.1045

Max Act: 0.0865. Min Act: -0.0000

Data Index: 8420 (C4 (Web Text))

Max Activating Token Index: 773

Text #15

Max Range: 0.1045. Min Range: -0.1045

Max Act: 0.0908. Min Act: -0.0000

Data Index: 409904 (C4 (Web Text))

Max Activating Token Index: 664

Text #16

Max Range: 0.1045. Min Range: -0.1045

Max Act: 0.0898. Min Act: -0.0000

Data Index: 1168502 (C4 (Web Text))

Max Activating Token Index: 947

Text #17

Max Range: 0.1045. Min Range: -0.1045

Max Act: 0.0864. Min Act: -0.0000

Data Index: 1038889 (C4 (Web Text))

Max Activating Token Index: 630

Text #18

Max Range: 0.1045. Min Range: -0.1045

Max Act: 0.0867. Min Act: -0.0000

Data Index: 1612122 (Python Code)

Max Activating Token Index: 497

Text #19

Max Range: 0.1045. Min Range: -0.1045

Max Act: 0.0881. Min Act: -0.0000

Data Index: 544812 (C4 (Web Text))

Max Activating Token Index: 705