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

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

Neuron 3766 in Layer 5

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.1309. Min Range: -0.1309

Max Act: 0.1309. Min Act: -0.0000

Data Index: 42236 (C4 (Web Text))

Max Activating Token Index: 77

Text #1

Max Range: 0.1309. Min Range: -0.1309

Max Act: 0.1164. Min Act: -0.0000

Data Index: 1427969 (Python Code)

Max Activating Token Index: 213

Text #2

Max Range: 0.1309. Min Range: -0.1309

Max Act: 0.1111. Min Act: -0.0000

Data Index: 336826 (C4 (Web Text))

Max Activating Token Index: 885

Text #3

Max Range: 0.1309. Min Range: -0.1309

Max Act: 0.0938. Min Act: -0.0000

Data Index: 647076 (C4 (Web Text))

Max Activating Token Index: 246

Text #4

Max Range: 0.1309. Min Range: -0.1309

Max Act: 0.0940. Min Act: -0.0000

Data Index: 274623 (C4 (Web Text))

Max Activating Token Index: 339

Text #5

Max Range: 0.1309. Min Range: -0.1309

Max Act: 0.0977. Min Act: -0.0000

Data Index: 64688 (C4 (Web Text))

Max Activating Token Index: 216

Text #6

Max Range: 0.1309. Min Range: -0.1309

Max Act: 0.0922. Min Act: -0.0000

Data Index: 389930 (C4 (Web Text))

Max Activating Token Index: 1018

Text #7

Max Range: 0.1309. Min Range: -0.1309

Max Act: 0.0928. Min Act: -0.0000

Data Index: 110325 (C4 (Web Text))

Max Activating Token Index: 576

Text #8

Max Range: 0.1309. Min Range: -0.1309

Max Act: 0.0989. Min Act: -0.0000

Data Index: 496373 (C4 (Web Text))

Max Activating Token Index: 905

Text #9

Max Range: 0.1309. Min Range: -0.1309

Max Act: 0.0887. Min Act: -0.0000

Data Index: 1555192 (Python Code)

Max Activating Token Index: 775

Text #10

Max Range: 0.1309. Min Range: -0.1309

Max Act: 0.0904. Min Act: -0.0000

Data Index: 288915 (C4 (Web Text))

Max Activating Token Index: 98

Text #11

Max Range: 0.1309. Min Range: -0.1309

Max Act: 0.0858. Min Act: -0.0000

Data Index: 1583217 (Python Code)

Max Activating Token Index: 303

Text #12

Max Range: 0.1309. Min Range: -0.1309

Max Act: 0.0912. Min Act: -0.0000

Data Index: 83721 (C4 (Web Text))

Max Activating Token Index: 244

Text #13

Max Range: 0.1309. Min Range: -0.1309

Max Act: 0.0826. Min Act: -0.0000

Data Index: 790781 (C4 (Web Text))

Max Activating Token Index: 629

Text #14

Max Range: 0.1309. Min Range: -0.1309

Max Act: 0.0833. Min Act: -0.0000

Data Index: 23760 (C4 (Web Text))

Max Activating Token Index: 118

Text #15

Max Range: 0.1309. Min Range: -0.1309

Max Act: 0.0817. Min Act: -0.0001

Data Index: 1019472 (C4 (Web Text))

Max Activating Token Index: 37

Text #16

Max Range: 0.1309. Min Range: -0.1309

Max Act: 0.0785. Min Act: -0.0001

Data Index: 1228732 (C4 (Web Text))

Max Activating Token Index: 941

Text #17

Max Range: 0.1309. Min Range: -0.1309

Max Act: 0.0807. Min Act: -0.0000

Data Index: 1419127 (Python Code)

Max Activating Token Index: 851

Text #18

Max Range: 0.1309. Min Range: -0.1309

Max Act: 0.0759. Min Act: -0.0000

Data Index: 231646 (C4 (Web Text))

Max Activating Token Index: 219

Text #19

Max Range: 0.1309. Min Range: -0.1309

Max Act: 0.0749. Min Act: -0.0001

Data Index: 720861 (C4 (Web Text))

Max Activating Token Index: 828