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

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

Neuron 186 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.0805. Min Range: -0.0805

Max Act: 0.0805. Min Act: -0.0005

Data Index: 11 (C4 (Web Text))

Max Activating Token Index: 7

Text #1

Max Range: 0.0805. Min Range: -0.0805

Max Act: 0.0676. Min Act: -0.0005

Data Index: 1114034 (C4 (Web Text))

Max Activating Token Index: 20

Text #2

Max Range: 0.0805. Min Range: -0.0805

Max Act: 0.0635. Min Act: -0.0005

Data Index: 1180781 (C4 (Web Text))

Max Activating Token Index: 525

Text #3

Max Range: 0.0805. Min Range: -0.0805

Max Act: 0.0591. Min Act: -0.0005

Data Index: 1021026 (C4 (Web Text))

Max Activating Token Index: 810

Text #4

Max Range: 0.0805. Min Range: -0.0805

Max Act: 0.0619. Min Act: -0.0005

Data Index: 216096 (C4 (Web Text))

Max Activating Token Index: 928

Text #5

Max Range: 0.0805. Min Range: -0.0805

Max Act: 0.0585. Min Act: -0.0005

Data Index: 892617 (C4 (Web Text))

Max Activating Token Index: 565

Text #6

Max Range: 0.0805. Min Range: -0.0805

Max Act: 0.0573. Min Act: -0.0005

Data Index: 696832 (C4 (Web Text))

Max Activating Token Index: 451

Text #7

Max Range: 0.0805. Min Range: -0.0805

Max Act: 0.0582. Min Act: -0.0005

Data Index: 202748 (C4 (Web Text))

Max Activating Token Index: 123

Text #8

Max Range: 0.0805. Min Range: -0.0805

Max Act: 0.0566. Min Act: -0.0005

Data Index: 1101942 (C4 (Web Text))

Max Activating Token Index: 755

Text #9

Max Range: 0.0805. Min Range: -0.0805

Max Act: 0.0583. Min Act: -0.0005

Data Index: 416576 (C4 (Web Text))

Max Activating Token Index: 424

Text #10

Max Range: 0.0805. Min Range: -0.0805

Max Act: 0.0641. Min Act: -0.0005

Data Index: 999341 (C4 (Web Text))

Max Activating Token Index: 695

Text #11

Max Range: 0.0805. Min Range: -0.0805

Max Act: 0.0578. Min Act: -0.0005

Data Index: 1366443 (Python Code)

Max Activating Token Index: 866

Text #12

Max Range: 0.0805. Min Range: -0.0805

Max Act: 0.0591. Min Act: -0.0005

Data Index: 638306 (C4 (Web Text))

Max Activating Token Index: 684

Text #13

Max Range: 0.0805. Min Range: -0.0805

Max Act: 0.0595. Min Act: -0.0005

Data Index: 1642965 (Python Code)

Max Activating Token Index: 1023

Text #14

Max Range: 0.0805. Min Range: -0.0805

Max Act: 0.0586. Min Act: -0.0005

Data Index: 1343400 (C4 (Web Text))

Max Activating Token Index: 38

Text #15

Max Range: 0.0805. Min Range: -0.0805

Max Act: 0.0503. Min Act: -0.0005

Data Index: 1201964 (C4 (Web Text))

Max Activating Token Index: 855

Text #16

Max Range: 0.0805. Min Range: -0.0805

Max Act: 0.0517. Min Act: -0.0005

Data Index: 1523198 (Python Code)

Max Activating Token Index: 957

Text #17

Max Range: 0.0805. Min Range: -0.0805

Max Act: 0.0556. Min Act: -0.0005

Data Index: 1104783 (C4 (Web Text))

Max Activating Token Index: 825

Text #18

Max Range: 0.0805. Min Range: -0.0805

Max Act: 0.0568. Min Act: -0.0005

Data Index: 1143536 (C4 (Web Text))

Max Activating Token Index: 478

Text #19

Max Range: 0.0805. Min Range: -0.0805

Max Act: 0.0604. Min Act: -0.0005

Data Index: 1398308 (Python Code)

Max Activating Token Index: 635