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

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

Neuron 4763 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: 2.6324. Min Range: -2.6324

Max Act: 2.6324. Min Act: -0.0000

Data Index: 1512697 (Python Code)

Max Activating Token Index: 330

Text #1

Max Range: 2.6324. Min Range: -2.6324

Max Act: 2.5165. Min Act: -0.0000

Data Index: 1008469 (C4 (Web Text))

Max Activating Token Index: 480

Text #2

Max Range: 2.6324. Min Range: -2.6324

Max Act: 1.9914. Min Act: -0.0000

Data Index: 257318 (C4 (Web Text))

Max Activating Token Index: 402

Text #3

Max Range: 2.6324. Min Range: -2.6324

Max Act: 1.9634. Min Act: -0.0000

Data Index: 1475067 (Python Code)

Max Activating Token Index: 90

Text #4

Max Range: 2.6324. Min Range: -2.6324

Max Act: 1.9112. Min Act: -0.0000

Data Index: 169133 (C4 (Web Text))

Max Activating Token Index: 179

Text #5

Max Range: 2.6324. Min Range: -2.6324

Max Act: 1.7946. Min Act: -0.0000

Data Index: 1612737 (Python Code)

Max Activating Token Index: 662

Text #6

Max Range: 2.6324. Min Range: -2.6324

Max Act: 1.8777. Min Act: -0.0000

Data Index: 1353986 (C4 (Web Text))

Max Activating Token Index: 501

Text #7

Max Range: 2.6324. Min Range: -2.6324

Max Act: 1.7628. Min Act: -0.0000

Data Index: 921398 (C4 (Web Text))

Max Activating Token Index: 362

Text #8

Max Range: 2.6324. Min Range: -2.6324

Max Act: 1.7509. Min Act: -0.0000

Data Index: 9706 (C4 (Web Text))

Max Activating Token Index: 792

Text #9

Max Range: 2.6324. Min Range: -2.6324

Max Act: 1.8461. Min Act: -0.0000

Data Index: 189388 (C4 (Web Text))

Max Activating Token Index: 277

Text #10

Max Range: 2.6324. Min Range: -2.6324

Max Act: 1.8395. Min Act: -0.0000

Data Index: 1583803 (Python Code)

Max Activating Token Index: 508

Text #11

Max Range: 2.6324. Min Range: -2.6324

Max Act: 1.9356. Min Act: -0.0000

Data Index: 1372480 (Python Code)

Max Activating Token Index: 371

Text #12

Max Range: 2.6324. Min Range: -2.6324

Max Act: 1.8248. Min Act: -0.0000

Data Index: 1488896 (Python Code)

Max Activating Token Index: 791

Text #13

Max Range: 2.6324. Min Range: -2.6324

Max Act: 1.7954. Min Act: -0.0000

Data Index: 869505 (C4 (Web Text))

Max Activating Token Index: 877

Text #14

Max Range: 2.6324. Min Range: -2.6324

Max Act: 1.7888. Min Act: -0.0000

Data Index: 1204085 (C4 (Web Text))

Max Activating Token Index: 439

Text #15

Max Range: 2.6324. Min Range: -2.6324

Max Act: 1.8029. Min Act: -0.0000

Data Index: 1464444 (Python Code)

Max Activating Token Index: 192

Text #16

Max Range: 2.6324. Min Range: -2.6324

Max Act: 1.8784. Min Act: -0.0000

Data Index: 1627989 (Python Code)

Max Activating Token Index: 993

Text #17

Max Range: 2.6324. Min Range: -2.6324

Max Act: 1.6723. Min Act: -0.0000

Data Index: 1094144 (C4 (Web Text))

Max Activating Token Index: 428

Text #18

Max Range: 2.6324. Min Range: -2.6324

Max Act: 1.6648. Min Act: -0.0000

Data Index: 1560310 (Python Code)

Max Activating Token Index: 752

Text #19

Max Range: 2.6324. Min Range: -2.6324

Max Act: 1.6597. Min Act: -0.0000

Data Index: 1376281 (Python Code)

Max Activating Token Index: 660