Model: SoLU Model: 3 Layers, 2048 Neurons per Layer

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

Neuron 1296 in Layer 0

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-3l')

Text #0

Max Range: 0.0101. Min Range: -0.0101

Max Act: 0.0101. Min Act: -0.0006

Data Index: 147329 (C4 (Web Text))

Max Activating Token Index: 983

Text #1

Max Range: 0.0101. Min Range: -0.0101

Max Act: 0.0097. Min Act: -0.0006

Data Index: 421455 (C4 (Web Text))

Max Activating Token Index: 706

Text #2

Max Range: 0.0101. Min Range: -0.0101

Max Act: 0.0086. Min Act: -0.0006

Data Index: 1072747 (C4 (Web Text))

Max Activating Token Index: 108

Text #3

Max Range: 0.0101. Min Range: -0.0101

Max Act: 0.0084. Min Act: -0.0006

Data Index: 695876 (C4 (Web Text))

Max Activating Token Index: 123

Text #4

Max Range: 0.0101. Min Range: -0.0101

Max Act: 0.0076. Min Act: -0.0006

Data Index: 96795 (C4 (Web Text))

Max Activating Token Index: 794

Text #5

Max Range: 0.0101. Min Range: -0.0101

Max Act: 0.0073. Min Act: -0.0005

Data Index: 1436409 (Python Code)

Max Activating Token Index: 1023

Text #6

Max Range: 0.0101. Min Range: -0.0101

Max Act: 0.0071. Min Act: -0.0006

Data Index: 1165772 (C4 (Web Text))

Max Activating Token Index: 383

Text #7

Max Range: 0.0101. Min Range: -0.0101

Max Act: 0.0072. Min Act: -0.0006

Data Index: 1024415 (C4 (Web Text))

Max Activating Token Index: 114

Text #8

Max Range: 0.0101. Min Range: -0.0101

Max Act: 0.0070. Min Act: -0.0005

Data Index: 1461905 (Python Code)

Max Activating Token Index: 40

Text #9

Max Range: 0.0101. Min Range: -0.0101

Max Act: 0.0070. Min Act: -0.0006

Data Index: 1225335 (C4 (Web Text))

Max Activating Token Index: 584

Text #10

Max Range: 0.0101. Min Range: -0.0101

Max Act: 0.0069. Min Act: -0.0006

Data Index: 1304930 (C4 (Web Text))

Max Activating Token Index: 500

Text #11

Max Range: 0.0101. Min Range: -0.0101

Max Act: 0.0072. Min Act: -0.0006

Data Index: 776343 (C4 (Web Text))

Max Activating Token Index: 853

Text #12

Max Range: 0.0101. Min Range: -0.0101

Max Act: 0.0068. Min Act: -0.0006

Data Index: 1140661 (C4 (Web Text))

Max Activating Token Index: 27

Text #13

Max Range: 0.0101. Min Range: -0.0101

Max Act: 0.0068. Min Act: -0.0005

Data Index: 1646491 (Python Code)

Max Activating Token Index: 251

Text #14

Max Range: 0.0101. Min Range: -0.0101

Max Act: 0.0067. Min Act: -0.0006

Data Index: 409506 (C4 (Web Text))

Max Activating Token Index: 605

Text #15

Max Range: 0.0101. Min Range: -0.0101

Max Act: 0.0069. Min Act: -0.0005

Data Index: 1499773 (Python Code)

Max Activating Token Index: 59

Text #16

Max Range: 0.0101. Min Range: -0.0101

Max Act: 0.0066. Min Act: -0.0005

Data Index: 1488339 (Python Code)

Max Activating Token Index: 869

Text #17

Max Range: 0.0101. Min Range: -0.0101

Max Act: 0.0068. Min Act: -0.0006

Data Index: 803424 (C4 (Web Text))

Max Activating Token Index: 293

Text #18

Max Range: 0.0101. Min Range: -0.0101

Max Act: 0.0068. Min Act: -0.0006

Data Index: 1478183 (Python Code)

Max Activating Token Index: 126

Text #19

Max Range: 0.0101. Min Range: -0.0101

Max Act: 0.0067. Min Act: -0.0006

Data Index: 1322045 (C4 (Web Text))

Max Activating Token Index: 227