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

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

Neuron 483 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 5.0664. Min Range: -5.0664

Max Act: 5.0664. Min Act: -0.1700

Data Index: 1145212 (C4 (Web Text))

Max Activating Token Index: 363

Text #1

Max Range: 5.0664. Min Range: -5.0664

Max Act: 5.0352. Min Act: -0.1700

Data Index: 295465 (C4 (Web Text))

Max Activating Token Index: 975

Text #2

Max Range: 5.0664. Min Range: -5.0664

Max Act: 4.9727. Min Act: -0.1700

Data Index: 610527 (C4 (Web Text))

Max Activating Token Index: 708

Text #3

Max Range: 5.0664. Min Range: -5.0664

Max Act: 4.9102. Min Act: -0.1700

Data Index: 469220 (C4 (Web Text))

Max Activating Token Index: 260

Text #4

Max Range: 5.0664. Min Range: -5.0664

Max Act: 4.8164. Min Act: -0.1700

Data Index: 384762 (C4 (Web Text))

Max Activating Token Index: 35

Text #5

Max Range: 5.0664. Min Range: -5.0664

Max Act: 4.8164. Min Act: -0.1700

Data Index: 509253 (C4 (Web Text))

Max Activating Token Index: 46

Text #6

Max Range: 5.0664. Min Range: -5.0664

Max Act: 4.6914. Min Act: -0.1700

Data Index: 1320701 (C4 (Web Text))

Max Activating Token Index: 506

Text #7

Max Range: 5.0664. Min Range: -5.0664

Max Act: 4.6914. Min Act: -0.1700

Data Index: 414670 (C4 (Web Text))

Max Activating Token Index: 354

Text #8

Max Range: 5.0664. Min Range: -5.0664

Max Act: 4.6602. Min Act: -0.1700

Data Index: 1187894 (C4 (Web Text))

Max Activating Token Index: 11

Text #9

Max Range: 5.0664. Min Range: -5.0664

Max Act: 4.6914. Min Act: -0.1700

Data Index: 648641 (C4 (Web Text))

Max Activating Token Index: 27

Text #10

Max Range: 5.0664. Min Range: -5.0664

Max Act: 4.5977. Min Act: -0.1700

Data Index: 845575 (C4 (Web Text))

Max Activating Token Index: 99

Text #11

Max Range: 5.0664. Min Range: -5.0664

Max Act: 4.6289. Min Act: -0.1700

Data Index: 1317340 (C4 (Web Text))

Max Activating Token Index: 22

Text #12

Max Range: 5.0664. Min Range: -5.0664

Max Act: 4.6602. Min Act: -0.1700

Data Index: 869863 (C4 (Web Text))

Max Activating Token Index: 111

Text #13

Max Range: 5.0664. Min Range: -5.0664

Max Act: 4.5664. Min Act: -0.1700

Data Index: 9199 (C4 (Web Text))

Max Activating Token Index: 173

Text #14

Max Range: 5.0664. Min Range: -5.0664

Max Act: 4.5977. Min Act: -0.1700

Data Index: 1616131 (Python Code)

Max Activating Token Index: 330

Text #15

Max Range: 5.0664. Min Range: -5.0664

Max Act: 4.5977. Min Act: -0.1700

Data Index: 1049148 (C4 (Web Text))

Max Activating Token Index: 528

Text #16

Max Range: 5.0664. Min Range: -5.0664

Max Act: 4.5039. Min Act: -0.1700

Data Index: 735243 (C4 (Web Text))

Max Activating Token Index: 552

Text #17

Max Range: 5.0664. Min Range: -5.0664

Max Act: 4.5977. Min Act: -0.1700

Data Index: 1344499 (C4 (Web Text))

Max Activating Token Index: 982

Text #18

Max Range: 5.0664. Min Range: -5.0664

Max Act: 4.5352. Min Act: -0.1700

Data Index: 444460 (C4 (Web Text))

Max Activating Token Index: 916

Text #19

Max Range: 5.0664. Min Range: -5.0664

Max Act: 4.6289. Min Act: -0.1700

Data Index: 1268162 (C4 (Web Text))

Max Activating Token Index: 256