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

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

Neuron 4484 in Layer 7

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.8278. Min Range: -0.8278

Max Act: 0.7748. Min Act: -0.0000

Data Index: 369084 (C4 (Web Text))

Max Activating Token Index: 182

Text #1

Max Range: 0.8278. Min Range: -0.8278

Max Act: 0.8278. Min Act: -0.0000

Data Index: 740718 (C4 (Web Text))

Max Activating Token Index: 934

Text #2

Max Range: 0.8278. Min Range: -0.8278

Max Act: 0.7292. Min Act: -0.0000

Data Index: 702191 (C4 (Web Text))

Max Activating Token Index: 587

Text #3

Max Range: 0.8278. Min Range: -0.8278

Max Act: 0.7330. Min Act: -0.0000

Data Index: 846323 (C4 (Web Text))

Max Activating Token Index: 546

Text #4

Max Range: 0.8278. Min Range: -0.8278

Max Act: 0.7211. Min Act: -0.0000

Data Index: 1490471 (Python Code)

Max Activating Token Index: 973

Text #5

Max Range: 0.8278. Min Range: -0.8278

Max Act: 0.7360. Min Act: -0.0000

Data Index: 283147 (C4 (Web Text))

Max Activating Token Index: 703

Text #6

Max Range: 0.8278. Min Range: -0.8278

Max Act: 0.7508. Min Act: -0.0000

Data Index: 385145 (C4 (Web Text))

Max Activating Token Index: 711

Text #7

Max Range: 0.8278. Min Range: -0.8278

Max Act: 0.7219. Min Act: -0.0000

Data Index: 1569454 (Python Code)

Max Activating Token Index: 526

Text #8

Max Range: 0.8278. Min Range: -0.8278

Max Act: 0.6471. Min Act: -0.0000

Data Index: 968201 (C4 (Web Text))

Max Activating Token Index: 695

Text #9

Max Range: 0.8278. Min Range: -0.8278

Max Act: 0.7113. Min Act: -0.0000

Data Index: 622465 (C4 (Web Text))

Max Activating Token Index: 243

Text #10

Max Range: 0.8278. Min Range: -0.8278

Max Act: 0.6292. Min Act: -0.0000

Data Index: 862715 (C4 (Web Text))

Max Activating Token Index: 240

Text #11

Max Range: 0.8278. Min Range: -0.8278

Max Act: 0.6943. Min Act: -0.0000

Data Index: 603915 (C4 (Web Text))

Max Activating Token Index: 225

Text #12

Max Range: 0.8278. Min Range: -0.8278

Max Act: 0.7178. Min Act: -0.0000

Data Index: 462707 (C4 (Web Text))

Max Activating Token Index: 865

Text #13

Max Range: 0.8278. Min Range: -0.8278

Max Act: 0.6856. Min Act: -0.0000

Data Index: 421595 (C4 (Web Text))

Max Activating Token Index: 135

Text #14

Max Range: 0.8278. Min Range: -0.8278

Max Act: 0.6388. Min Act: -0.0000

Data Index: 714120 (C4 (Web Text))

Max Activating Token Index: 255

Text #15

Max Range: 0.8278. Min Range: -0.8278

Max Act: 0.6063. Min Act: -0.0000

Data Index: 708102 (C4 (Web Text))

Max Activating Token Index: 612

Text #16

Max Range: 0.8278. Min Range: -0.8278

Max Act: 0.5943. Min Act: -0.0000

Data Index: 1209520 (C4 (Web Text))

Max Activating Token Index: 990

Text #17

Max Range: 0.8278. Min Range: -0.8278

Max Act: 0.5951. Min Act: -0.0000

Data Index: 1630565 (Python Code)

Max Activating Token Index: 135

Text #18

Max Range: 0.8278. Min Range: -0.8278

Max Act: 0.5665. Min Act: -0.0000

Data Index: 1532029 (Python Code)

Max Activating Token Index: 903

Text #19

Max Range: 0.8278. Min Range: -0.8278

Max Act: 0.5509. Min Act: -0.0000

Data Index: 899556 (C4 (Web Text))

Max Activating Token Index: 646