Model: SoLU Model: 4 Layers, 2048 Neurons per Layer

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

Neuron 1915 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-4l')

Text #0

Max Range: 0.2572. Min Range: -0.2572

Max Act: 0.2572. Min Act: -0.0002

Data Index: 357438 (C4 (Web Text))

Max Activating Token Index: 243

Text #1

Max Range: 0.2572. Min Range: -0.2572

Max Act: 0.2559. Min Act: -0.0002

Data Index: 1401337 (Python Code)

Max Activating Token Index: 248

Text #2

Max Range: 0.2572. Min Range: -0.2572

Max Act: 0.2280. Min Act: -0.0002

Data Index: 1281909 (C4 (Web Text))

Max Activating Token Index: 284

Text #3

Max Range: 0.2572. Min Range: -0.2572

Max Act: 0.2106. Min Act: -0.0001

Data Index: 35681 (C4 (Web Text))

Max Activating Token Index: 834

Text #4

Max Range: 0.2572. Min Range: -0.2572

Max Act: 0.2283. Min Act: -0.0002

Data Index: 700516 (C4 (Web Text))

Max Activating Token Index: 380

Text #5

Max Range: 0.2572. Min Range: -0.2572

Max Act: 0.2222. Min Act: -0.0002

Data Index: 445379 (C4 (Web Text))

Max Activating Token Index: 716

Text #6

Max Range: 0.2572. Min Range: -0.2572

Max Act: 0.2165. Min Act: -0.0002

Data Index: 387360 (C4 (Web Text))

Max Activating Token Index: 288

Text #7

Max Range: 0.2572. Min Range: -0.2572

Max Act: 0.2140. Min Act: -0.0002

Data Index: 1454201 (Python Code)

Max Activating Token Index: 358

Text #8

Max Range: 0.2572. Min Range: -0.2572

Max Act: 0.2143. Min Act: -0.0002

Data Index: 5852 (C4 (Web Text))

Max Activating Token Index: 748

Text #9

Max Range: 0.2572. Min Range: -0.2572

Max Act: 0.1989. Min Act: -0.0002

Data Index: 766716 (C4 (Web Text))

Max Activating Token Index: 577

Text #10

Max Range: 0.2572. Min Range: -0.2572

Max Act: 0.2238. Min Act: -0.0002

Data Index: 1158986 (C4 (Web Text))

Max Activating Token Index: 946

Text #11

Max Range: 0.2572. Min Range: -0.2572

Max Act: 0.1653. Min Act: -0.0002

Data Index: 159444 (C4 (Web Text))

Max Activating Token Index: 365

Text #12

Max Range: 0.2572. Min Range: -0.2572

Max Act: 0.1966. Min Act: -0.0002

Data Index: 703912 (C4 (Web Text))

Max Activating Token Index: 962

Text #13

Max Range: 0.2572. Min Range: -0.2572

Max Act: 0.1823. Min Act: -0.0001

Data Index: 968612 (C4 (Web Text))

Max Activating Token Index: 392

Text #14

Max Range: 0.2572. Min Range: -0.2572

Max Act: 0.1809. Min Act: -0.0002

Data Index: 171336 (C4 (Web Text))

Max Activating Token Index: 346

Text #15

Max Range: 0.2572. Min Range: -0.2572

Max Act: 0.1851. Min Act: -0.0002

Data Index: 1028987 (C4 (Web Text))

Max Activating Token Index: 7

Text #16

Max Range: 0.2572. Min Range: -0.2572

Max Act: 0.1833. Min Act: -0.0002

Data Index: 1496524 (Python Code)

Max Activating Token Index: 923

Text #17

Max Range: 0.2572. Min Range: -0.2572

Max Act: 0.1751. Min Act: -0.0002

Data Index: 9217 (C4 (Web Text))

Max Activating Token Index: 647

Text #18

Max Range: 0.2572. Min Range: -0.2572

Max Act: 0.1768. Min Act: -0.0002

Data Index: 527894 (C4 (Web Text))

Max Activating Token Index: 191

Text #19

Max Range: 0.2572. Min Range: -0.2572

Max Act: 0.1671. Min Act: -0.0002

Data Index: 47399 (C4 (Web Text))

Max Activating Token Index: 442