Model: SoLU Model: 12 Layers, 6144 Neurons per Layer

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

Neuron 5197 in Layer 8

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-12l')

Text #0

Max Range: 2.5895. Min Range: -2.5895

Max Act: 2.5895. Min Act: -0.0000

Data Index: 654273 (C4 (Web Text))

Max Activating Token Index: 1019

Text #1

Max Range: 2.5895. Min Range: -2.5895

Max Act: 1.4414. Min Act: -0.0000

Data Index: 1319036 (C4 (Web Text))

Max Activating Token Index: 912

Text #2

Max Range: 2.5895. Min Range: -2.5895

Max Act: 1.1695. Min Act: -0.0000

Data Index: 1589988 (Python Code)

Max Activating Token Index: 748

Text #3

Max Range: 2.5895. Min Range: -2.5895

Max Act: 1.2260. Min Act: -0.0000

Data Index: 1310883 (C4 (Web Text))

Max Activating Token Index: 948

Text #4

Max Range: 2.5895. Min Range: -2.5895

Max Act: 1.2133. Min Act: -0.0000

Data Index: 1595742 (Python Code)

Max Activating Token Index: 688

Text #5

Max Range: 2.5895. Min Range: -2.5895

Max Act: 1.2061. Min Act: -0.0000

Data Index: 187539 (C4 (Web Text))

Max Activating Token Index: 90

Text #6

Max Range: 2.5895. Min Range: -2.5895

Max Act: 1.0294. Min Act: -0.0000

Data Index: 412854 (C4 (Web Text))

Max Activating Token Index: 457

Text #7

Max Range: 2.5895. Min Range: -2.5895

Max Act: 0.9107. Min Act: -0.0000

Data Index: 839889 (C4 (Web Text))

Max Activating Token Index: 766

Text #8

Max Range: 2.5895. Min Range: -2.5895

Max Act: 0.9787. Min Act: -0.0000

Data Index: 1519786 (Python Code)

Max Activating Token Index: 53

Text #9

Max Range: 2.5895. Min Range: -2.5895

Max Act: 0.8615. Min Act: -0.0000

Data Index: 1629967 (Python Code)

Max Activating Token Index: 600

Text #10

Max Range: 2.5895. Min Range: -2.5895

Max Act: 0.9616. Min Act: -0.0000

Data Index: 180497 (C4 (Web Text))

Max Activating Token Index: 497

Text #11

Max Range: 2.5895. Min Range: -2.5895

Max Act: 0.9356. Min Act: -0.0000

Data Index: 1601669 (Python Code)

Max Activating Token Index: 578

Text #12

Max Range: 2.5895. Min Range: -2.5895

Max Act: 0.8310. Min Act: -0.0000

Data Index: 1112752 (C4 (Web Text))

Max Activating Token Index: 907

Text #13

Max Range: 2.5895. Min Range: -2.5895

Max Act: 0.7992. Min Act: -0.0000

Data Index: 1491149 (Python Code)

Max Activating Token Index: 85

Text #14

Max Range: 2.5895. Min Range: -2.5895

Max Act: 0.8683. Min Act: -0.0000

Data Index: 572866 (C4 (Web Text))

Max Activating Token Index: 643

Text #15

Max Range: 2.5895. Min Range: -2.5895

Max Act: 0.8382. Min Act: -0.0000

Data Index: 451806 (C4 (Web Text))

Max Activating Token Index: 593

Text #16

Max Range: 2.5895. Min Range: -2.5895

Max Act: 0.8180. Min Act: -0.0000

Data Index: 1447035 (Python Code)

Max Activating Token Index: 30

Text #17

Max Range: 2.5895. Min Range: -2.5895

Max Act: 0.8640. Min Act: -0.0000

Data Index: 502600 (C4 (Web Text))

Max Activating Token Index: 951

Text #18

Max Range: 2.5895. Min Range: -2.5895

Max Act: 0.9259. Min Act: -0.0000

Data Index: 155160 (C4 (Web Text))

Max Activating Token Index: 680

Text #19

Max Range: 2.5895. Min Range: -2.5895

Max Act: 0.9168. Min Act: -0.0000

Data Index: 201783 (C4 (Web Text))

Max Activating Token Index: 521