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

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

Neuron 420 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: 0.0845. Min Range: -0.0845

Max Act: 0.0833. Min Act: -0.0002

Data Index: 760834 (C4 (Web Text))

Max Activating Token Index: 93

Text #1

Max Range: 0.0845. Min Range: -0.0845

Max Act: 0.0845. Min Act: -0.0002

Data Index: 1041893 (C4 (Web Text))

Max Activating Token Index: 692

Text #2

Max Range: 0.0845. Min Range: -0.0845

Max Act: 0.0811. Min Act: -0.0002

Data Index: 1091700 (C4 (Web Text))

Max Activating Token Index: 61

Text #3

Max Range: 0.0845. Min Range: -0.0845

Max Act: 0.0817. Min Act: -0.0002

Data Index: 533556 (C4 (Web Text))

Max Activating Token Index: 58

Text #4

Max Range: 0.0845. Min Range: -0.0845

Max Act: 0.0805. Min Act: -0.0002

Data Index: 18633 (C4 (Web Text))

Max Activating Token Index: 714

Text #5

Max Range: 0.0845. Min Range: -0.0845

Max Act: 0.0809. Min Act: -0.0002

Data Index: 310807 (C4 (Web Text))

Max Activating Token Index: 508

Text #6

Max Range: 0.0845. Min Range: -0.0845

Max Act: 0.0784. Min Act: -0.0002

Data Index: 637666 (C4 (Web Text))

Max Activating Token Index: 182

Text #7

Max Range: 0.0845. Min Range: -0.0845

Max Act: 0.0809. Min Act: -0.0002

Data Index: 204896 (C4 (Web Text))

Max Activating Token Index: 755

Text #8

Max Range: 0.0845. Min Range: -0.0845

Max Act: 0.0777. Min Act: -0.0002

Data Index: 481436 (C4 (Web Text))

Max Activating Token Index: 406

Text #9

Max Range: 0.0845. Min Range: -0.0845

Max Act: 0.0789. Min Act: -0.0002

Data Index: 431353 (C4 (Web Text))

Max Activating Token Index: 65

Text #10

Max Range: 0.0845. Min Range: -0.0845

Max Act: 0.0776. Min Act: -0.0002

Data Index: 1245052 (C4 (Web Text))

Max Activating Token Index: 55

Text #11

Max Range: 0.0845. Min Range: -0.0845

Max Act: 0.0760. Min Act: -0.0002

Data Index: 110431 (C4 (Web Text))

Max Activating Token Index: 348

Text #12

Max Range: 0.0845. Min Range: -0.0845

Max Act: 0.0761. Min Act: -0.0002

Data Index: 624752 (C4 (Web Text))

Max Activating Token Index: 247

Text #13

Max Range: 0.0845. Min Range: -0.0845

Max Act: 0.0770. Min Act: -0.0002

Data Index: 1141222 (C4 (Web Text))

Max Activating Token Index: 412

Text #14

Max Range: 0.0845. Min Range: -0.0845

Max Act: 0.0759. Min Act: -0.0001

Data Index: 335613 (C4 (Web Text))

Max Activating Token Index: 329

Text #15

Max Range: 0.0845. Min Range: -0.0845

Max Act: 0.0761. Min Act: -0.0002

Data Index: 445507 (C4 (Web Text))

Max Activating Token Index: 814

Text #16

Max Range: 0.0845. Min Range: -0.0845

Max Act: 0.0744. Min Act: -0.0002

Data Index: 119740 (C4 (Web Text))

Max Activating Token Index: 808

Text #17

Max Range: 0.0845. Min Range: -0.0845

Max Act: 0.0751. Min Act: -0.0002

Data Index: 1513693 (Python Code)

Max Activating Token Index: 906

Text #18

Max Range: 0.0845. Min Range: -0.0845

Max Act: 0.0745. Min Act: -0.0002

Data Index: 87138 (C4 (Web Text))

Max Activating Token Index: 717

Text #19

Max Range: 0.0845. Min Range: -0.0845

Max Act: 0.0725. Min Act: -0.0002

Data Index: 565715 (C4 (Web Text))

Max Activating Token Index: 266